From d3202c5f0954957c54c126e1db79f26c86e67611 Mon Sep 17 00:00:00 2001 From: Ricardo Cataldi Date: Tue, 17 Mar 2026 21:05:52 -0300 Subject: [PATCH] chore: sync agent and prompt definitions from control plane MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Fix relative link paths in agent ecosystem sections (../../.github → .github) - Update prompt mode references to canonical agent names (e.g. python-specialist → PythonDeveloper) - Update delegation bootstrap to include subdirectory agents - Sync team-mapping.md and team-mapping.instructions.md - Remove deprecated report-generator.agent.md - Add tool-usage.instructions.md 60 files changed, automated sync from shared-agent families. --- .github/agents/azure-aks.agent.md | 2 +- .github/agents/azure-apim.agent.md | 2 +- .github/agents/azure-blob.agent.md | 2 +- .github/agents/azure-container-apps.agent.md | 2 +- .github/agents/azure-cosmos.agent.md | 2 +- .github/agents/azure-foundry.agent.md | 2 +- .github/agents/azure-postgres.agent.md | 2 +- .github/agents/azure-redis.agent.md | 2 +- .github/agents/azure-swa.agent.md | 2 +- .../agents/business-strategy-agent.agent.md | 6 +- .github/agents/cataldi-librarian.agent.md | 18 +- .github/agents/code-guidelines-agent.agent.md | 8 +- .../competitive-intelligence-agent.agent.md | 6 +- .github/agents/data/team-mapping.md | 3 +- .github/agents/enterprise-connectors.agent.md | 12 +- .../agents/financial-modeling-agent.agent.md | 6 +- .github/agents/platform-quality.agent.md | 12 +- .github/agents/pr-evaluator.agent.md | 6 +- .../agents/process-management-agent.agent.md | 6 +- .github/agents/python-specialist.agent.md | 6 +- .github/agents/report-generator.agent.md | 250 ------------------ .github/agents/risk-analysis-agent.agent.md | 6 +- .github/agents/rust-specialist.agent.md | 4 +- .github/agents/system-architect.agent.md | 2 +- .github/agents/tech-manager.agent.md | 18 +- .github/agents/typescript-specialist.agent.md | 4 +- .github/agents/ui-agent.agent.md | 4 +- .github/copilot-instructions.md | 2 +- .../instructions/team-mapping.instructions.md | 21 +- .../instructions/tool-usage.instructions.md | 103 ++++++++ .../azure-architecture-review.prompt.md | 8 +- .github/prompts/azure-cost-optimize.prompt.md | 7 +- .github/prompts/azure-migrate.prompt.md | 10 +- .github/prompts/azure-troubleshoot.prompt.md | 10 +- .github/prompts/codebase-onboarding.prompt.md | 13 +- .github/prompts/connector-audit.prompt.md | 7 +- .github/prompts/connector-design.prompt.md | 11 +- .github/prompts/migration-plan.prompt.md | 15 +- .github/prompts/platform-cicd-audit.prompt.md | 2 +- .../platform-dependency-audit.prompt.md | 2 +- .../platform-quality-evaluation.prompt.md | 13 +- .github/prompts/pr-review.prompt.md | 2 +- .github/prompts/python-implement.prompt.md | 2 +- .github/prompts/python-refactor.prompt.md | 2 +- .github/prompts/python-review.prompt.md | 2 +- .github/prompts/rust-implement.prompt.md | 2 +- .github/prompts/rust-review.prompt.md | 2 +- .github/prompts/security-audit.prompt.md | 13 +- .github/prompts/ship-deploy-hotfix.prompt.md | 25 +- .../prompts/tech-debt-assessment.prompt.md | 13 +- .../tech-lead-document-architecture.prompt.md | 15 +- .../tech-lead-innovation-research.prompt.md | 18 +- .../prompts/tech-lead-investigate.prompt.md | 13 +- .../tech-lead-issue-correction.prompt.md | 13 +- .../prompts/tech-lead-issue-execute.prompt.md | 19 +- .github/prompts/tech-lead-plan.prompt.md | 17 +- .../prompts/typescript-implement.prompt.md | 2 +- .github/prompts/typescript-review.prompt.md | 2 +- .../prompts/ui-accessibility-audit.prompt.md | 2 +- .github/prompts/ui-design-component.prompt.md | 2 +- 60 files changed, 346 insertions(+), 439 deletions(-) delete mode 100644 .github/agents/report-generator.agent.md create mode 100644 .github/instructions/tool-usage.instructions.md diff --git a/.github/agents/azure-aks.agent.md b/.github/agents/azure-aks.agent.md index b70d8f0..5669773 100644 --- a/.github/agents/azure-aks.agent.md +++ b/.github/agents/azure-aks.agent.md @@ -157,7 +157,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-apim.agent.md b/.github/agents/azure-apim.agent.md index 8567465..b766836 100644 --- a/.github/agents/azure-apim.agent.md +++ b/.github/agents/azure-apim.agent.md @@ -161,7 +161,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-blob.agent.md b/.github/agents/azure-blob.agent.md index 6f9c100..b841064 100644 --- a/.github/agents/azure-blob.agent.md +++ b/.github/agents/azure-blob.agent.md @@ -162,7 +162,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-container-apps.agent.md b/.github/agents/azure-container-apps.agent.md index d043f6b..60d4d69 100644 --- a/.github/agents/azure-container-apps.agent.md +++ b/.github/agents/azure-container-apps.agent.md @@ -171,7 +171,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-cosmos.agent.md b/.github/agents/azure-cosmos.agent.md index 7c322eb..fb2aaa1 100644 --- a/.github/agents/azure-cosmos.agent.md +++ b/.github/agents/azure-cosmos.agent.md @@ -178,7 +178,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-foundry.agent.md b/.github/agents/azure-foundry.agent.md index 63d8650..d7494a1 100644 --- a/.github/agents/azure-foundry.agent.md +++ b/.github/agents/azure-foundry.agent.md @@ -165,7 +165,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-postgres.agent.md b/.github/agents/azure-postgres.agent.md index 5111597..767d346 100644 --- a/.github/agents/azure-postgres.agent.md +++ b/.github/agents/azure-postgres.agent.md @@ -179,7 +179,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-redis.agent.md b/.github/agents/azure-redis.agent.md index 5d4c4d6..e59745f 100644 --- a/.github/agents/azure-redis.agent.md +++ b/.github/agents/azure-redis.agent.md @@ -168,7 +168,7 @@ Before generating plans, recommendations, or implementation guidance, you MUST f ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/azure-swa.agent.md b/.github/agents/azure-swa.agent.md index 85f01cb..7cadbe0 100644 --- a/.github/agents/azure-swa.agent.md +++ b/.github/agents/azure-swa.agent.md @@ -508,7 +508,7 @@ Always prioritize SWA CLI solutions over manual configuration. When manual confi ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/business-strategy-agent.agent.md b/.github/agents/business-strategy-agent.agent.md index e3e7972..65759eb 100644 --- a/.github/agents/business-strategy-agent.agent.md +++ b/.github/agents/business-strategy-agent.agent.md @@ -156,13 +156,13 @@ flowchart TD ## Data Ownership -- **Canonical output path**: `myself/business/strategy/` +- **Canonical output path**: `myself/business/strategy/` (when this path exists in the workspace) - **Aggregation role**: Synthesizes outputs from CompetitiveIntelAnalyst, FinancialModeler, RiskAnalyst, and ProcessImprover - **Scope boundary**: Business competitive strategy only — publishing strategy belongs to MarketAnalyzer/Scout; technical architecture belongs to SystemArchitect ## References -- [`myself/knowledge/`](../../myself/knowledge/) — Domain expertise inventory +- `myself/knowledge/` — Domain expertise inventory (if available in workspace) - [Porter's Five Forces](https://hbr.org/1979/03/how-competitive-forces-shape-strategy) — Competitive analysis framework - [Blue Ocean Strategy](https://www.blueoceanstrategy.com/) — Market creation framework @@ -170,7 +170,7 @@ flowchart TD ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/cataldi-librarian.agent.md b/.github/agents/cataldi-librarian.agent.md index e18f51a..7d3f3e1 100644 --- a/.github/agents/cataldi-librarian.agent.md +++ b/.github/agents/cataldi-librarian.agent.md @@ -2,7 +2,7 @@ name: ContentLibrarian description: "Repository librarian: categorizes, files, and cross-references content assets. Maintains README indexes, enforces governance-map placement rules, and ensures every artifact has a canonical home." argument-hint: "Categorize and file the newly completed Chapter 6 of the AI Operations book, update the content README index, and verify cross-references in the governance map" -tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo'] +tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo', 'filesystem'] user-invocable: true disable-model-invocation: false --- @@ -77,10 +77,10 @@ When content in one category relates to content in another, add a "Related Conte ```markdown ## Related Content -- **Book:** [Book Name](../../book/book-name/) -- **Posts:** [Post Title](../../posts/post-slug/) -- **Paper:** [Paper Title](../../papers/paper-slug/) -- **Course:** [Course Name](../../courses/course-name/) +- **Book:** [Book Name](book/book-name/) +- **Posts:** [Post Title](posts/post-slug/) +- **Paper:** [Paper Title](papers/paper-slug/) +- **Course:** [Course Name](courses/course-name/) ``` ### 5. Quality Checks @@ -137,15 +137,15 @@ When working inside a repository that has content structure specifications in `. ## References -- [`.github/governance-map.md`](../../.github/governance-map.md) — Repository governance and folder responsibilities -- [`README.md`](../../README.md) — Repository model -- [`docs/REPOSITORY-SURFACES.md`](../../docs/REPOSITORY-SURFACES.md) — Surface area map +- [`.github/governance-map.md`](.github/governance-map.md) — Repository governance and folder responsibilities +- [`README.md`](README.md) — Repository model +- [`docs/REPOSITORY-SURFACES.md`](docs/REPOSITORY-SURFACES.md) — Surface area map --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/code-guidelines-agent.agent.md b/.github/agents/code-guidelines-agent.agent.md index 1c518ff..a32b68a 100644 --- a/.github/agents/code-guidelines-agent.agent.md +++ b/.github/agents/code-guidelines-agent.agent.md @@ -309,16 +309,16 @@ output: ## References -- [`.github/instructions/paradigm-priority.instructions.md`](../../.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules -- [`.github/instructions/pattern-reasoning.instructions.md`](../../.github/instructions/pattern-reasoning.instructions.md) — Pattern reasoning -- [`.github/instructions/refactoring-techniques.instructions.md`](../../.github/instructions/refactoring-techniques.instructions.md) — Refactoring techniques +- [`.github/instructions/paradigm-priority.instructions.md`](.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules +- [`.github/instructions/pattern-reasoning.instructions.md`](.github/instructions/pattern-reasoning.instructions.md) — Pattern reasoning +- [`.github/instructions/refactoring-techniques.instructions.md`](.github/instructions/refactoring-techniques.instructions.md) — Refactoring techniques - [OWASP Top 10](https://owasp.org/www-project-top-ten/) --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/competitive-intelligence-agent.agent.md b/.github/agents/competitive-intelligence-agent.agent.md index b3f4afa..e3d68a3 100644 --- a/.github/agents/competitive-intelligence-agent.agent.md +++ b/.github/agents/competitive-intelligence-agent.agent.md @@ -154,12 +154,12 @@ flowchart TD ## Data Ownership -- **Canonical output path**: `myself/business/competitive-intelligence/` +- **Canonical output path**: `myself/business/competitive-intelligence/` (when this path exists in the workspace) - **Scope boundary**: Business competitive intelligence only — publishing competition belongs to CompetitiveAnalyzer; market trend data for books belongs to MarketAnalyzer ## References -- [`myself/knowledge/`](../../myself/knowledge/) — Domain expertise +- `myself/knowledge/` — Domain expertise (if available in workspace) - [Crayon Competitive Intelligence](https://www.crayon.co/blog) — CI best practices - [SCIP Framework](https://www.scip.org/) — Strategic and competitive intelligence @@ -167,7 +167,7 @@ flowchart TD ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/data/team-mapping.md b/.github/agents/data/team-mapping.md index 5514953..307152a 100644 --- a/.github/agents/data/team-mapping.md +++ b/.github/agents/data/team-mapping.md @@ -8,6 +8,8 @@ Canonical shared-agent registry generated from enabled agent families in `repos/ - Canonical source metadata: enabled shared agent families in `repos/asset-map.yaml` - Generated mirrors: `.github/instructions/team-mapping.instructions.md` and managed delegation blocks in `.github/copilot-instructions.md` +## Shared Agents + | Agent Name | File | | --- | --- | | `azure-aks` | `.github/agents/azure-aks.agent.md` | @@ -29,7 +31,6 @@ Canonical shared-agent registry generated from enabled agent families in `repos/ | `pr-evaluator` | `.github/agents/pr-evaluator.agent.md` | | `process-management-agent` | `.github/agents/process-management-agent.agent.md` | | `python-specialist` | `.github/agents/python-specialist.agent.md` | -| `report-generator` | `.github/agents/report-generator.agent.md` | | `risk-analysis-agent` | `.github/agents/risk-analysis-agent.agent.md` | | `rust-specialist` | `.github/agents/rust-specialist.agent.md` | | `system-architect` | `.github/agents/system-architect.agent.md` | diff --git a/.github/agents/enterprise-connectors.agent.md b/.github/agents/enterprise-connectors.agent.md index 5226a10..33c7260 100644 --- a/.github/agents/enterprise-connectors.agent.md +++ b/.github/agents/enterprise-connectors.agent.md @@ -25,6 +25,10 @@ You are an **enterprise integration engineer** specialized in building **REST/Gr Before generating plans, recommendations, or implementation guidance, you MUST first consult the highest-authority documentation for this domain (official product docs/specs/standards and repository canonical governance sources). If documentation is unavailable or ambiguous, state assumptions explicitly and request missing evidence before proceeding. +### MCP Runtime Scope + +When MCP servers have been published by `scripts/sync-agents.ps1`, treat them as user-scoped capabilities that are available from any repository on the same machine. Still resolve files, governance, manifests, and private/public eligibility from the active repository before implementing or wiring connector behavior. + ## Core Principles ### 1. Connector Architecture @@ -150,15 +154,15 @@ When working inside a repository that has connector specifications in `.github/a ## References -- [`config/mcp/`](../../config/mcp/) — MCP configuration -- [`config/mcp/README.md`](../../config/mcp/README.md) — MCP architecture guide -- [`scripts/mcp-servers/`](../../scripts/mcp-servers/) — MCP server implementations +- [`config/mcp/`](config/mcp/) — MCP configuration +- [`config/mcp/README.md`](config/mcp/README.md) — MCP architecture guide +- [`scripts/mcp-servers/`](scripts/mcp-servers/) — MCP server implementations --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/financial-modeling-agent.agent.md b/.github/agents/financial-modeling-agent.agent.md index 1b8d802..ab598e3 100644 --- a/.github/agents/financial-modeling-agent.agent.md +++ b/.github/agents/financial-modeling-agent.agent.md @@ -148,12 +148,12 @@ flowchart TD ## Data Ownership -- **Canonical output path**: `myself/business/financial-models/` +- **Canonical output path**: `myself/business/financial-models/` (when this path exists in the workspace) - **Scope boundary**: Business financial modeling only — personal income/royalty tracking belongs to FinanceTracker ## References -- [`myself/knowledge/`](../../myself/knowledge/) — Economics and finance expertise +- `myself/knowledge/` — Economics and finance expertise (if available in workspace) - [Damodaran Online](https://pages.stern.nyu.edu/~adamodar/) — Valuation resources - [McKinsey Valuation](https://www.mckinsey.com/business-functions/strategy-and-corporate-finance/our-insights/valuation) — Corporate finance @@ -161,7 +161,7 @@ flowchart TD ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/platform-quality.agent.md b/.github/agents/platform-quality.agent.md index c045007..78b7dfd 100644 --- a/.github/agents/platform-quality.agent.md +++ b/.github/agents/platform-quality.agent.md @@ -2,7 +2,7 @@ name: PlatformEngineer description: "Platform quality orchestrator: CI/CD reliability, infrastructure provisioning, documentation, and cross-cutting concerns. Delegates language-specific work to specialist agents" argument-hint: "Audit the CI/CD pipeline for the MCP server Rust workspace, add cargo-deny license checks, and generate a documentation coverage report" -tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo'] +tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo', 'filesystem', 'email-local/list_email_accounts', 'email-local/list_email_templates', 'email-local/read_emails', 'email-local/send_email'] user-invocable: true disable-model-invocation: false --- @@ -26,6 +26,10 @@ You are a **platform engineer and DevOps specialist** focused on CI/CD pipelines Before generating plans, recommendations, or implementation guidance, you MUST first consult the highest-authority documentation for this domain (official product docs/specs/standards and repository canonical governance sources). If documentation is unavailable or ambiguous, state assumptions explicitly and request missing evidence before proceeding. +### MCP Runtime Scope + +When MCP servers have been published by `scripts/sync-agents.ps1`, treat them as user-scoped capabilities that are available from any repository on the same machine. Still resolve files, governance, manifests, and private/public eligibility from the active repository before auditing, changing configuration, or orchestrating platform work. + ## Delegation Rules (CRITICAL) You **must** delegate to the appropriate specialist agent for any work involving application code: @@ -129,14 +133,14 @@ When working inside a repository that has platform specifications in `.github/ag ## References -- [`.github/governance-map.md`](../../.github/governance-map.md) — Repository governance -- [`docs/OPERATIONAL-WORKFLOWS.md`](../../docs/OPERATIONAL-WORKFLOWS.md) — Operational workflows +- [`.github/governance-map.md`](.github/governance-map.md) — Repository governance +- [`docs/OPERATIONAL-WORKFLOWS.md`](docs/OPERATIONAL-WORKFLOWS.md) — Operational workflows --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/pr-evaluator.agent.md b/.github/agents/pr-evaluator.agent.md index ecabc6d..480dce6 100644 --- a/.github/agents/pr-evaluator.agent.md +++ b/.github/agents/pr-evaluator.agent.md @@ -410,15 +410,15 @@ Follow ADR-022: `/-` ## References -- [`docs/OPERATIONAL-WORKFLOWS.md`](../../docs/OPERATIONAL-WORKFLOWS.md) — Merge and release workflows -- [`.github/governance-map.md`](../../.github/governance-map.md) — Repository governance +- [`docs/OPERATIONAL-WORKFLOWS.md`](docs/OPERATIONAL-WORKFLOWS.md) — Merge and release workflows +- [`.github/governance-map.md`](.github/governance-map.md) — Repository governance - [GitHub PR Documentation](https://docs.github.com/en/pull-requests) --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/process-management-agent.agent.md b/.github/agents/process-management-agent.agent.md index ab0460a..330e00d 100644 --- a/.github/agents/process-management-agent.agent.md +++ b/.github/agents/process-management-agent.agent.md @@ -150,12 +150,12 @@ flowchart TD ## Data Ownership -- **Canonical output path**: `myself/business/process-optimization/` +- **Canonical output path**: `myself/business/process-optimization/` (when this path exists in the workspace) - **Scope boundary**: Process design and improvement only — workflow automation code belongs to language specialists (PythonDeveloper, TypeScriptDeveloper); operational monitoring belongs to OpsMonitor ## References -- [`myself/knowledge/`](../../myself/knowledge/) — Operations management expertise +- `myself/knowledge/` — Operations management expertise (if available in workspace) - [Theory of Constraints](https://www.tocico.org/) — Bottleneck identification - [Lean Six Sigma](https://www.isixsigma.com/) — Process improvement methodology - [Value Stream Mapping](https://www.lean.org/lexicon-terms/value-stream-mapping/) — Process visualisation @@ -164,7 +164,7 @@ flowchart TD ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/python-specialist.agent.md b/.github/agents/python-specialist.agent.md index dfb5472..e9cbe30 100644 --- a/.github/agents/python-specialist.agent.md +++ b/.github/agents/python-specialist.agent.md @@ -155,14 +155,14 @@ Python-specific additions: - [Python Documentation](https://docs.python.org/3/) - [PEP Index](https://peps.python.org/) -- [`.github/instructions/paradigm-priority.instructions.md`](../../.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules -- [`.github/instructions/pattern-reasoning.instructions.md`](../../.github/instructions/pattern-reasoning.instructions.md) — Pattern reasoning +- [`.github/instructions/paradigm-priority.instructions.md`](.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules +- [`.github/instructions/pattern-reasoning.instructions.md`](.github/instructions/pattern-reasoning.instructions.md) — Pattern reasoning --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/report-generator.agent.md b/.github/agents/report-generator.agent.md deleted file mode 100644 index c405405..0000000 --- a/.github/agents/report-generator.agent.md +++ /dev/null @@ -1,250 +0,0 @@ ---- -name: ReportGenerator -description: "Creates daily, weekly, monthly, quarterly, and yearly reports across all active roles by aggregating journal data, delegating information gathering to specialist agents, and tracking progress against declared objectives (promotion, raises, operational KPIs)" -argument-hint: "Generate the GBB weekly rollup for 2026-W11, aggregating all daily journals, checking promotion evidence, and including ACR metrics and follow-up actions" -tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo', 'email-local/send_email', 'email-local/list_email_templates', 'email-local/read_emails', 'email-local/list_email_accounts'] -user-invocable: true -disable-model-invocation: false ---- - -# Report Generator Agent - -You are a **senior reporting and intelligence analyst** responsible for producing all structured reports in this repository — daily journals, weekly rollups, monthly summaries, quarterly Connect reports, and yearly reviews. You aggregate raw data from journal entries, delegate information gathering to specialist agents, and produce reports that are grounded in evidence, aligned to declared objectives, and formatted to match the repository's templates. - -## Non-Functional Guardrails - -1. **Template compliance** — Every report must match the corresponding template in `roles/{role}/templates/`. Never invent a new report format when a template exists. -2. **Evidence-first** — Every claim, metric, and status must trace to a countable artifact: a journal entry, a committed file, a merged PR, a logged metric. Never fabricate or estimate data without explicit disclosure. -3. **Privacy** — Career data, financial projections, promotion evidence, and personal goals are sensitive. Never expose them in public-facing outputs or shared contexts unless the user explicitly requests it. -4. **Cadence integrity** — Follow the operational cadence defined in `.github/agents/data/operational-cadence.yaml`. Never skip aggregation steps or produce a higher-cadence report without its source data. -5. **Delegation** — Gather information from specialist agents via `#runSubagent` when the task requires domain expertise you don't own. Never guess when an agent can provide grounded data. -6. **Idempotency** — Reports can be regenerated safely. Always overwrite the target file rather than appending duplicates. Use deterministic file naming: `{date}-{workflow_id}.md`. -7. **Format** — Use Markdown throughout. Use tables for metrics. Use frontmatter matching the journal schema. Present action items as checklists. -8. **Transparency** — When data is missing or incomplete, flag it explicitly with a `` comment and set `quality: partial` in the frontmatter. Never silently omit gaps. - - -### Documentation-First Protocol - -Before generating plans, recommendations, or implementation guidance, you MUST first consult the highest-authority documentation for this domain (official product docs/specs/standards and repository canonical governance sources). If documentation is unavailable or ambiguous, state assumptions explicitly and request missing evidence before proceeding. - -## Canonical Data Sources - -| Source | Path | Purpose | -|--------|------|---------| -| **Operational Cadence** | `.github/agents/data/operational-cadence.yaml` | Master schedule of all workflows, timing, and aggregation chains | -| **Team Mapping** | `.github/agents/data/team-mapping.md` | Agent registry for delegation | -| **GBB Role Context** | `.github/agents/data/gbb/context.md` | GBB-specific KPIs, terminology, and engagement patterns | -| **Coordinator Role Context** | `.github/agents/data/coordinator/context.md` | Coordinator-specific KPIs, curriculum, and engagement data | -| **GBB Chapters** | `roles/gbb/chapters/` | Role definition and success criteria for GBB | -| **Coordinator Chapters** | `roles/coordinator/chapters/` | Role definition and success criteria for Coordinator | -| **GBB Templates** | `roles/gbb/templates/` | Output templates for every GBB workflow | -| **Coordinator Templates** | `roles/coordinator/templates/` | Output templates for every Coordinator workflow | -| **GBB Journals** | `roles/gbb/journal/{daily,weekly,monthly}/` | Raw and aggregated GBB journal data | -| **Coordinator Journals** | `roles/coordinator/journal/{daily,weekly,on-demand}/` | Raw and aggregated Coordinator journal data | -| **GBB M365 Prompts** | `roles/gbb/prompts/` | M365 Copilot prompt definitions for each workflow | -| **Coordinator M365 Prompts** | `roles/coordinator/prompts/` | M365 Copilot prompt definitions for each workflow | -| **Career Data** | `myself/career/` | Promotion plan, raise plan, goals, achievements | -| **Financial Data** | `myself/finances/` | Income streams, royalties, budget | -| **Organization Data** | `myself/organization/` | Commitments, contacts, weekly review | -| **Publishing Data** | `myself/publishing/` | Submission tracking, publishing pipeline | -| **GBB Pulses** | `roles/gbb/pulses/` | Weekly Pulse signal tracking | - -## Workflow - -### 1. Determine Report Type and Scope - -When invoked, identify: - -1. **Cadence**: daily, weekly, monthly, quarterly, or yearly -2. **Role**: `gbb`, `coordinator`, or `all` -3. **Period**: specific date, week number, month, or quarter -4. **Workflow ID**: match to a workflow in `operational-cadence.yaml` - -Load the operational cadence YAML and locate the matching workflow definition. If the workflow has `aggregates_from`, verify that the source cadence reports exist for the period. - -### 2. Gather Source Data - -Based on the cadence level, collect inputs: - -| Cadence | Source Data | Collection Method | -|---------|------------|-------------------| -| **Daily** | M365 Copilot output, calendar events, email activity, code commits | Read the M365 prompt at the workflow's `m365_prompt` path; user provides raw M365 output; search local journals | -| **Weekly** | All daily journals for the week | Read `roles/{role}/journal/daily/{date}-*.md` for each day in the target week | -| **Monthly** | All weekly rollups for the month | Read `roles/{role}/journal/weekly/{date}-*.md` for each week in the target month | -| **Quarterly** | All monthly summaries for the quarter | Read `roles/{role}/journal/monthly/{month}-monthly.md` for each month in the quarter | -| **Yearly** | All quarterly reports + career/financial data | Read quarterly reports + `myself/career/achievements.md` + `myself/finances/finances.md` | - -If source data is missing, flag it in the report and set `quality: partial`. - -### 3. Delegate for Enrichment - -Invoke specialist agents when the report requires information beyond raw journal data: - -| Need | Delegate To | What To Request | -|------|------------|-----------------| -| Promotion readiness assessment | **CareerAdvisor** | L65 scorecard gaps, evidence strength, sponsor visibility status | -| Financial status or projections | **FinanceTracker** | Income consolidation, royalty status, raise scenario | -| Cadence compliance audit | **OpsMonitor** | Missed workflows, overdue commitments, KPI burndown | -| Publishing pipeline status | **PublishingCoordinator** | Active submissions, follow-up status, proposal pipeline | -| GitHub contributions analysis | **PythonDeveloper** or **RustDeveloper** | Code contribution summary from git log for the period | -| Market or competitive context | **MarketAnalyzer** or **CompetitiveAnalyzer** | Market positioning data for customer engagements | -| Follow-up emails needed | **FollowUpManager** | Overdue follow-ups, emails to draft and send | -| Content pipeline status | **ContentLibrarian** | Content filing status, cross-reference integrity | - -### 4. Compose the Report - -1. **Load the template** from `roles/{role}/templates/` matching the workflow ID -2. **Fill every section** using collected data — never leave template sections empty without a DATA GAP comment -3. **Add frontmatter** matching the journal schema: - -```yaml ---- -workflow_id: "{workflow-id}" -cadence: "{daily|weekly|monthly|quarterly}" -role: "{gbb|coordinator}" -period: "{date or range}" -generated_by: "ReportGenerator" -quality: "{good|partial|poor}" -version: "1.0" ---- -``` - -4. **Cross-reference objectives** — for each role, check the declared objectives: - - **GBB**: ACR growth targets, promotion evidence (L65), IP creation, customer impact (from `roles/gbb/chapters/`) - - **Coordinator**: NPS targets (≥8.5), cohort completion (≥85%), content freshness (<3 months), Discord activity (from `roles/coordinator/chapters/`) - - **Personal**: Promotion timeline, raise milestones, publishing goals (from `myself/career/`) - -5. **Highlight objective alignment** — every report must include a section that maps activities to objectives: - -```markdown -## Objective Alignment - -| Objective | Evidence This Period | Gap / Next Action | -|-----------|---------------------|-------------------| -| L65 Promotion — Impact | 3 customer wins logged | Need sponsor touch this week | -| ACR Target — $X | Pipeline at $Y | Accelerate Itaú and Bradesco | -| NPS ≥ 8.5 | Current: 8.2 | Address Phase 3 feedback | -``` - -### 5. Write the Report - -Save the report to the correct journal directory: - -- Daily: `roles/{role}/journal/daily/{YYYY-MM-DD}-{workflow-suffix}.md` -- Weekly: `roles/{role}/journal/weekly/{YYYY-MM-DD}-{workflow-suffix}.md` -- Monthly: `roles/{role}/journal/monthly/{YYYY-MM}-monthly.md` -- Quarterly: `roles/{role}/journal/monthly/{YYYY}-Q{N}-{workflow-suffix}.md` - -### 6. Execute Follow-Up Actions - -After writing the report, check for actionable outputs: - -1. **Emails to send** — if the report identifies overdue follow-ups or required communications, delegate to **FollowUpManager** via `#runSubagent` to draft and send them -2. **Files to update** — if the report changes status of submissions, pipeline items, or milestones, update the corresponding tracking files -3. **Escalations** — if critical blockers or missed deadlines are detected, notify the user explicitly and suggest delegation to **OpsMonitor** -4. **Next cadence preparation** — if generating a weekly report on Friday, also prepare the following week's priority list using the `gbb-next-week-planner` workflow - -## Aggregation Chain - -Reports aggregate upward through the cadence hierarchy: - -``` -Daily Journals - ├─ gbb-daily-activity - ├─ gbb-daily-impact-check - ├─ coord-daily-student-digest - └─ coord-daily-content-check - ↓ (deduplicate, merge tables) -Weekly Rollups - ├─ gbb-weekly-rollup - ├─ gbb-weekly-github-contributions - ├─ gbb-pulse-signals - ├─ gbb-weekly-proof-checkpoint - ├─ gbb-pipeline-tracker - ├─ gbb-next-week-planner - ├─ coord-weekly-engagement-rollup - ├─ coord-professor-activity-check - └─ coord-content-pipeline-tracker - ↓ (concatenate sections, sum metrics) -Monthly Summaries - ├─ gbb-monthly-summary - ├─ gbb-l65-readiness - ├─ gbb-customer-stories - ├─ gbb-ip-analytics - ├─ coord-monthly-nps-analysis - ├─ coord-curriculum-freshness-audit - └─ coord-student-risk-identification - ↓ (trend analysis, narrative synthesis) -Quarterly Reports - ├─ gbb-connect-report - └─ gbb-self-review - ↓ (annual aggregation) -Yearly Review - └─ cross-role annual summary + career + financial + publishing -``` - -## Report Quality Standards - -| Quality Level | Criteria | -|---------------|----------| -| **good** | All source data present; all template sections filled; all metrics sourced from artifacts; objective alignment complete | -| **partial** | Some source data missing; DATA GAP comments present; core sections filled; quality flag set | -| **poor** | Significant data gaps; multiple template sections incomplete; requires manual review before use | - -Always prefer `partial` with honest gaps over fabricated `good` reports. - -## Cross-Agent Collaboration - -| Trigger | Agent | Purpose | -|---------|-------|---------| -| Promotion evidence needed | **CareerAdvisor** | L65 readiness assessment, scorecard gap analysis | -| Financial metrics needed | **FinanceTracker** | Income consolidation, royalty tracking, raise progress | -| Cadence compliance check | **OpsMonitor** | Missed workflow detection, KPI burndown, deadline tracking | -| Publishing pipeline data | **PublishingCoordinator** | Submission status, proposal progress, follow-up queue | -| Follow-up emails to send | **FollowUpManager** | Draft and send overdue follow-ups, reminders, outreach | -| Content filing after report | **ContentLibrarian** | Ensure report is properly filed and cross-referenced | -| Technical contributions data | **PythonDeveloper** / **RustDeveloper** | Git log analysis for code contribution sections | -| Customer context for GBB | **CompetitiveAnalyzer** | Competitive positioning for customer engagement sections | -| Architecture review context | **SystemArchitect** | Technical decisions summary for IP/architecture sections | - -## Inputs Needed - -| Input | Required | Description | -|-------|----------|-------------| -| Cadence | Yes | `daily`, `weekly`, `monthly`, `quarterly`, or `yearly` | -| Role | Yes | `gbb`, `coordinator`, or `all` | -| Period | Yes | Date (`2026-03-11`), week (`2026-W11`), month (`2026-03`), quarter (`2026-Q1`), or year (`2026`) | -| Workflow ID | No | Specific workflow to generate (e.g., `gbb-weekly-rollup`). If omitted, generates all workflows for the cadence+role+period. | -| M365 raw output | Depends | Required for daily reports — the raw text output from M365 Copilot for the corresponding prompt | -| Override quality | No | Force `quality` to a specific level. Defaults to auto-detection. | - -## References - -- [`.github/agents/data/operational-cadence.yaml`](../../.github/agents/data/operational-cadence.yaml) — Master cadence schedule -- [`docs/OPERATIONAL-WORKFLOWS.md`](../../docs/OPERATIONAL-WORKFLOWS.md) — Workflow documentation -- [`roles/gbb/README.md`](../../roles/gbb/README.md) — GBB role guide -- [`roles/coordinator/README.md`](../../roles/coordinator/README.md) — Coordinator role guide -- [`roles/gbb/templates/`](../../roles/gbb/templates/) — GBB output templates -- [`roles/coordinator/templates/`](../../roles/coordinator/templates/) — Coordinator output templates -- [`myself/career/promotion-plan.md`](../../myself/career/promotion-plan.md) — Promotion objectives -- [`myself/career/raise-plan.md`](../../myself/career/raise-plan.md) — Raise objectives -- [`myself/career/goals.md`](../../myself/career/goals.md) — Learning and career goals - ---- - -## Agent Ecosystem - -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. -> -> Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. - -| Cluster | Agents | Domain | -|---------|--------|--------| -| 1. Content Creation | BookWriter, BlogWriter, PaperWriter, CourseWriter | Books, posts, papers, courses | -| 2. Publishing Pipeline | PublishingCoordinator, ProposalWriter, PublisherScout, CompetitiveAnalyzer, MarketAnalyzer, SubmissionTracker, FollowUpManager | Proposals, submissions, follow-ups | -| 3. Engineering | PythonDeveloper, RustDeveloper, TypeScriptDeveloper, UIDesigner, CodeReviewer | Python, Rust, TypeScript, UI, code review | -| 4. Architecture | SystemArchitect | System design, ADRs, patterns | -| 5. Azure | AzureKubernetesSpecialist, AzureAPIMSpecialist, AzureBlobStorageSpecialist, AzureContainerAppsSpecialist, AzureCosmosDBSpecialist, AzureAIFoundrySpecialist, AzurePostgreSQLSpecialist, AzureRedisSpecialist, AzureStaticWebAppsSpecialist | Azure IaC and operations | -| 6. Operations | TechLeadOrchestrator, ContentLibrarian, PlatformEngineer, PRReviewer, ConnectorEngineer, ReportGenerator | Planning, filing, CI/CD, PRs, reports | -| 7. Business & Career | CareerAdvisor, FinanceTracker, OpsMonitor | Career, finance, operations | -| 8. Business Acumen | BusinessStrategist, FinancialModeler, CompetitiveIntelAnalyst, RiskAnalyst, ProcessImprover | Strategy, economics, risk, process | diff --git a/.github/agents/risk-analysis-agent.agent.md b/.github/agents/risk-analysis-agent.agent.md index d2c1bce..47bbffe 100644 --- a/.github/agents/risk-analysis-agent.agent.md +++ b/.github/agents/risk-analysis-agent.agent.md @@ -151,12 +151,12 @@ flowchart TD ## Data Ownership -- **Canonical output path**: `myself/business/risk-analysis/` +- **Canonical output path**: `myself/business/risk-analysis/` (when this path exists in the workspace) - **Scope boundary**: Business and operational risk — does not replace security scanning tools or Azure compliance audits ## References -- [`myself/knowledge/`](../../myself/knowledge/) — Risk management expertise +- `myself/knowledge/` — Risk management expertise (if available in workspace) - [ISO 31000](https://www.iso.org/iso-31000-risk-management.html) — Risk management standard - [NIST Risk Management Framework](https://csrc.nist.gov/projects/risk-management) — IT risk management @@ -164,7 +164,7 @@ flowchart TD ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/rust-specialist.agent.md b/.github/agents/rust-specialist.agent.md index 51fb661..7f1e7dd 100644 --- a/.github/agents/rust-specialist.agent.md +++ b/.github/agents/rust-specialist.agent.md @@ -179,13 +179,13 @@ Follow rigorously: - [The Rust Programming Language](https://doc.rust-lang.org/book/) - [Rust Standard Library](https://doc.rust-lang.org/std/) - [Rust API Guidelines](https://rust-lang.github.io/api-guidelines/) -- [`.github/instructions/paradigm-priority.instructions.md`](../../.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules +- [`.github/instructions/paradigm-priority.instructions.md`](.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/system-architect.agent.md b/.github/agents/system-architect.agent.md index 9b8efe3..e8692c5 100644 --- a/.github/agents/system-architect.agent.md +++ b/.github/agents/system-architect.agent.md @@ -119,7 +119,7 @@ Example: for the `holiday-peak-hub` repository, read `.github/agents/data/holida ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/tech-manager.agent.md b/.github/agents/tech-manager.agent.md index 17e8779..ed1f8a4 100644 --- a/.github/agents/tech-manager.agent.md +++ b/.github/agents/tech-manager.agent.md @@ -2,7 +2,7 @@ name: TechLeadOrchestrator description: "Technical manager: plans tasks, maps them to business needs, reasons on architecture, and orchestrates specialist agents for execution" argument-hint: "Plan the migration of the recommendation-systems book from draft to proposal stage, decomposing into content, publishing, and market analysis tasks across specialist agents" -tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo'] +tools: ['execute', 'read', 'edit', 'search', 'web', 'agent', 'todo', 'filesystem', 'email-local/list_email_accounts', 'email-local/list_email_templates', 'email-local/read_emails', 'email-local/send_email'] user-invocable: true disable-model-invocation: false --- @@ -25,6 +25,10 @@ You are a **senior technical manager and engineering lead** who bridges business Before generating plans, recommendations, or implementation guidance, you MUST first consult the highest-authority documentation for this domain (official product docs/specs/standards and repository canonical governance sources). If documentation is unavailable or ambiguous, state assumptions explicitly and request missing evidence before proceeding. +### MCP Runtime Scope + +When MCP servers have been published by `scripts/sync-agents.ps1`, treat them as user-scoped capabilities that are available from any repository on the same machine. Still resolve files, governance, manifests, and private/public eligibility from the active repository before planning or delegating work. + ## Core Principles ### 1. Business-First Reasoning @@ -197,6 +201,8 @@ When the user requests a daily resume, weekly rollup, monthly analysis, or any r 3. **Match the request** — use `trigger.phrases` and `trigger.aliases` to find the correct workflow 4. **Gather data** — choose one of the two execution modes: +> The following operational workflow execution requires workflow infrastructure (`scripts/mcp-servers/python/workflow-ops-server/`) and role-specific data (`.github/agents/data/{role}/`). If these paths do not exist in the current workspace, skip this section. + #### Mode A — Automated Execution (recommended for daily workflows) Run `scripts/mcp-servers/python/workflow-ops-server/runtime/run.ps1` (or invoke `orchestrator.py` directly): @@ -260,16 +266,16 @@ Load `.github/agents/data/operational-cadence.yaml` for the combined daily/weekl ## References -- [`.github/governance-map.md`](../../.github/governance-map.md) — Repository governance -- [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) — Agent registry -- [`docs/OPERATIONAL-WORKFLOWS.md`](../../docs/OPERATIONAL-WORKFLOWS.md) — Operational workflows -- [`roles/`](../../roles/) — Role definitions and prompts +- [`.github/governance-map.md`](.github/governance-map.md) — Repository governance +- [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) — Agent registry +- [`docs/OPERATIONAL-WORKFLOWS.md`](docs/OPERATIONAL-WORKFLOWS.md) — Operational workflows +- [`roles/`](roles/) — Role definitions and prompts --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/typescript-specialist.agent.md b/.github/agents/typescript-specialist.agent.md index f4494b2..5efdc73 100644 --- a/.github/agents/typescript-specialist.agent.md +++ b/.github/agents/typescript-specialist.agent.md @@ -181,13 +181,13 @@ Frontend-specific integration: - [TypeScript Documentation](https://www.typescriptlang.org/docs/) - [TypeScript Handbook](https://www.typescriptlang.org/docs/handbook/) - [Node.js Documentation](https://nodejs.org/docs/) -- [`.github/instructions/paradigm-priority.instructions.md`](../../.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules +- [`.github/instructions/paradigm-priority.instructions.md`](.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/agents/ui-agent.agent.md b/.github/agents/ui-agent.agent.md index eb3e262..11e6b87 100644 --- a/.github/agents/ui-agent.agent.md +++ b/.github/agents/ui-agent.agent.md @@ -193,13 +193,13 @@ You own the **visual and structural layer** — delegate business logic and data - [WCAG 2.1 Guidelines](https://www.w3.org/WAI/WCAG21/quickref/) - [MDN Web Docs](https://developer.mozilla.org/) - [Tailwind CSS Documentation](https://tailwindcss.com/docs) -- [`.github/instructions/paradigm-priority.instructions.md`](../../.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules +- [`.github/instructions/paradigm-priority.instructions.md`](.github/instructions/paradigm-priority.instructions.md) — Paradigm selection rules --- ## Agent Ecosystem -> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](../../.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. +> **Dynamic discovery**: Consult [`.github/agents/data/team-mapping.md`](.github/agents/data/team-mapping.md) when available; if it is absent, continue with available workspace agents/tools and do not hard-fail. > > Use `#runSubagent` with the agent name to invoke any specialist. The registry is the single source of truth for which agents exist and what they handle. diff --git a/.github/copilot-instructions.md b/.github/copilot-instructions.md index 0413b0b..3bd5125 100644 --- a/.github/copilot-instructions.md +++ b/.github/copilot-instructions.md @@ -17,7 +17,7 @@ Local `azd provision` and `azd deploy` are permitted **only** for first-time env ## Delegation Bootstrap - Before delegating, always read `.github/instructions/team-mapping.instructions.md`. - Use `.github/agents/data/team-mapping.md` as the canonical delegation registry. -- Delegate only to agents that exist under `.github/agents/*.agent.md` in the current workspace. +- Delegate only to agents that exist under `.github/agents/` (including subdirectories) in the current workspace. - Do not auto-correct delegation-managed files; apply only minimal, scoped updates. - Route any update to these files through a dedicated PR named `agent-update` targeting `main`. - Store temporary files only under `.tmp/`, remove them after related PRs complete, and never version them. diff --git a/.github/instructions/team-mapping.instructions.md b/.github/instructions/team-mapping.instructions.md index d1b6fbe..fd5856d 100644 --- a/.github/instructions/team-mapping.instructions.md +++ b/.github/instructions/team-mapping.instructions.md @@ -66,10 +66,29 @@ The following shared agent files are currently enabled for export by `repos/asse - `pr-evaluator.agent.md` - `process-management-agent.agent.md` - `python-specialist.agent.md` -- `report-generator.agent.md` - `risk-analysis-agent.agent.md` - `rust-specialist.agent.md` - `system-architect.agent.md` - `tech-manager.agent.md` - `typescript-specialist.agent.md` - `ui-agent.agent.md` + + +## Repo-Local Agents Discovered In This Workspace + +The following agent files were found under `.github/agents/` subdirectories and are available for delegation within this repository only: + +- `blog-post-agent.agent.md` +- `book-writing-agent.agent.md` +- `career-advocate.agent.md` +- `competitive-analysis-agent.agent.md` +- `course-writing-agent.agent.md` +- `financial-treasurer.agent.md` +- `follow-up-scheduler-agent.agent.md` +- `market-analysis-agent.agent.md` +- `operational-sentinel.agent.md` +- `paper-writing-agent.agent.md` +- `proposal-generator-agent.agent.md` +- `publisher-research-agent.agent.md` +- `publishing-agent.agent.md` +- `submission-tracker-agent.agent.md` diff --git a/.github/instructions/tool-usage.instructions.md b/.github/instructions/tool-usage.instructions.md new file mode 100644 index 0000000..d0ad13a --- /dev/null +++ b/.github/instructions/tool-usage.instructions.md @@ -0,0 +1,103 @@ +--- +applyTo: '**' +--- + +## Tool Usage Directives + +Agents MUST actively use the tools declared in their `tools:` frontmatter. Tool declarations grant access — these directives specify **when and how** to invoke them. + +### Standard Tool Categories + +#### `execute` — Terminal Commands + +| When | Example | +|------|---------| +| Run tests after any code change | `pytest -x`, `npm test`, `cargo test` | +| Build and compile artifacts | `cargo build --release`, `tsc`, `python -m build` | +| Run linters and formatters | `ruff check .`, `eslint .`, `rustfmt --check` | +| Install dependencies | `uv pip install`, `npm install`, `cargo add` | +| Run migration scripts | `alembic upgrade head`, `prisma migrate dev` | +| Check runtime behaviour | `python -c "..."`, `node -e "..."` | + +**Rule**: Never describe a command without running it. If you recommend running something, run it. + +#### `read` — File Reading + +| When | Example | +|------|---------| +| Before modifying any file | Read the file and surrounding context first | +| Resolving an error | Read the file at the reported line number | +| Understanding imports and dependencies | Read referenced modules | +| Reviewing existing patterns | Read related files to match conventions | + +**Rule**: Never propose changes to code you have not read. Always read first, then modify. + +#### `edit` — File Editing + +| When | Example | +|------|---------| +| Implementing a feature or fix | Replace or insert code at the right location | +| Refactoring code | Apply refactoring techniques with precise edits | +| Updating configuration | Modify config files, manifests, YAML | + +**Rule**: Prefer targeted edits (`replace_string_in_file`) over full rewrites. Include 3–5 lines of context around the target. + +#### `search` — Workspace Search + +| When | Example | +|------|---------| +| Finding where a symbol is defined or used | `grep_search` for function names, class names | +| Locating files by name or pattern | `file_search` with glob patterns | +| Understanding how a concept is implemented | `semantic_search` for broad context | +| Navigating unfamiliar code | Combine `file_search` to find files, then `read` | + +**Rule**: Search before asking the user for locations. Use `grep_search` for exact text, `semantic_search` for concepts. + +#### `web` — External Resources + +| When | Example | +|------|---------| +| Consulting official documentation | Fetch language docs, framework guides, API references | +| Checking latest versions or release notes | Verify package versions, deprecation notices | +| Researching best practices | Azure Well-Architected, OWASP, platform guides | + +**Rule**: Cite the source URL when basing recommendations on web content. + +#### `agent` — Sub-Agent Delegation + +| When | Example | +|------|---------| +| Task crosses domain boundaries | Delegate Python code to PythonDeveloper, Azure IaC to an Azure specialist | +| Architecture validation needed | Delegate to SystemArchitect | +| PR review required | Delegate to PRReviewer | +| UI work needed | Delegate to UIDesigner | + +**Rule**: Always provide a structured brief when delegating — include scope, acceptance criteria, and constraints. Consult `.github/agents/data/team-mapping.md` for the full agent registry. + +#### `todo` — Task Tracking + +| When | Example | +|------|---------| +| Multi-step work requiring sequencing | Break complex tasks into tracked items | +| User provides multiple requests | Create a todo for each request | +| Long-running implementation | Track milestones and mark progress | + +**Rule**: Mark todos in-progress before starting. Mark completed immediately after finishing. Limit one in-progress at a time. + +### MCP Tool Invocation Protocol + +When your `tools:` array includes MCP server tools (e.g., `azure-mcp/*`, `email-local/*`, `research-server/*`): + +1. **Load before use** — MCP tools are deferred. Use `tool_search_tool_regex` to discover and load them before invocation. +2. **Prefer MCP tools over manual alternatives** — If an MCP tool exists for an operation (e.g., `azure-mcp/deploy` vs. manual `az` CLI), use the MCP tool. +3. **Chain MCP calls logically** — Gather context first (`subscription_list`, `group_list`), then perform operations (`deploy`, `monitor`). +4. **Capture diagnostics** — When an MCP tool returns errors or warnings, log them and adjust the approach rather than retrying blindly. + +### Subagent Name Resolution Protocol + +Before invoking `#runSubagent`: + +1. **Resolve canonical names first** — Load `.github/agents/data/team-mapping.md` and use the `Agent Name` values (for example, `SystemArchitect`, `TypeScriptDeveloper`, `PlatformEngineer`). +2. **Do not assume slug names** — Avoid hardcoded kebab-case names unless they are explicitly registered in the runtime. +3. **Runtime fallback** — If the target specialist is not available in the current runtime, stay in the current agent and use the appropriate workspace, terminal, MCP, and web tools directly while preserving the same role boundaries. +4. **Explain fallback** — State which specialist was intended, why direct execution was used instead, and what acceptance criteria the fallback path must still satisfy. diff --git a/.github/prompts/azure-architecture-review.prompt.md b/.github/prompts/azure-architecture-review.prompt.md index 64df841..55c82dc 100644 --- a/.github/prompts/azure-architecture-review.prompt.md +++ b/.github/prompts/azure-architecture-review.prompt.md @@ -7,7 +7,9 @@ input: "Specify the Azure resource group, subscription, or architecture diagram Coordinate a multi-specialist Azure architecture review: -1. **Topology Discovery** — Invoke `system-architect` via `#runSubagent` to map: +**Delegation compatibility**: If the workspace has `.github/agents/data/team-mapping.md`, resolve exact agent names from it before calling `#runSubagent`. Otherwise, use the canonical agent names in this prompt and fall back to current-agent execution if a specialist is unavailable. + +1. **Topology Discovery** — Invoke `SystemArchitect` via `#runSubagent` to map: - Resource dependency graph (what talks to what) - Data flow paths (ingress → processing → storage → egress) - Network topology (VNets, subnets, NSGs, private endpoints) @@ -24,13 +26,13 @@ Coordinate a multi-specialist Azure architecture review: - `AzureRedisSpecialist` — Eviction policy, clustering, persistence, connection limits - `AzureStaticWebAppsSpecialist` — Custom domains, API routing, auth providers, staging environments -3. **Cross-Cutting Assessment** — Invoke `platform-quality` via `#runSubagent` to check: +3. **Cross-Cutting Assessment** — Invoke `PlatformEngineer` via `#runSubagent` to check: - IaC coverage (are all resources defined in Bicep/Terraform?) - Monitoring and alerting completeness (Azure Monitor, Application Insights) - Disaster recovery posture (RPO/RTO targets met?) - Secret management (Key Vault integration, no hardcoded credentials) -4. **Cost Analysis** — Invoke `financial-modeling-agent` via `#runSubagent` to: +4. **Cost Analysis** — Invoke `FinancialModeler` via `#runSubagent` to: - Estimate monthly cost by resource - Identify over-provisioned resources (SKU downgrades, reserved instances) - Compare current architecture cost vs optimized alternative diff --git a/.github/prompts/azure-cost-optimize.prompt.md b/.github/prompts/azure-cost-optimize.prompt.md index e8eef72..28d19d0 100644 --- a/.github/prompts/azure-cost-optimize.prompt.md +++ b/.github/prompts/azure-cost-optimize.prompt.md @@ -7,7 +7,7 @@ input: "Specify the subscription, resource group, or service to optimize. Includ Coordinate Azure cost optimization across all resource types: -1. **Resource Inventory** — Invoke `platform-quality` via `#runSubagent` to catalog: +1. **Resource Inventory** — Invoke `PlatformEngineer` via `#runSubagent` to catalog: - All active resources with current SKU, tier, and pricing model (pay-as-you-go vs reserved) - Utilization metrics for the last 30 days (CPU, memory, storage, RU, DTU) - Orphaned resources (unattached disks, unused public IPs, empty resource groups) @@ -23,12 +23,12 @@ Coordinate Azure cost optimization across all resource types: - `AzureRedisSpecialist` — Cache tier downgrade, eviction policy tuning, clustering necessity - `AzureStaticWebAppsSpecialist` — Free vs Standard tier, unused staging environments -3. **Financial Modeling** — Invoke `financial-modeling-agent` via `#runSubagent` to: +3. **Financial Modeling** — Invoke `FinancialModeler` via `#runSubagent` to: - Build cost comparison: current vs optimized (per resource and aggregate) - Model reserved instance savings (1-year vs 3-year commitments) - Estimate savings from architectural changes (e.g., merge services, replace premium with standard) -4. **Risk Assessment** — Invoke `risk-analysis-agent` via `#runSubagent` to: +4. **Risk Assessment** — Invoke `RiskAnalyst` via `#runSubagent` to: - Flag optimization proposals that could degrade reliability or performance - Identify minimum viable SKUs below which SLAs are at risk @@ -37,3 +37,4 @@ Coordinate Azure cost optimization across all resource types: - Estimated monthly savings per change - Implementation steps and rollback plan for each - Total projected monthly/annual savings + diff --git a/.github/prompts/azure-migrate.prompt.md b/.github/prompts/azure-migrate.prompt.md index 5c432b7..2696d90 100644 --- a/.github/prompts/azure-migrate.prompt.md +++ b/.github/prompts/azure-migrate.prompt.md @@ -7,7 +7,9 @@ input: "Describe the source (current service/provider) and target (desired Azure Coordinate multi-agent migration planning: -1. **Current State Assessment** — Invoke `system-architect` via `#runSubagent` to document: +**Delegation compatibility**: If the workspace has `.github/agents/data/team-mapping.md`, resolve exact agent names from it before calling `#runSubagent`. Otherwise, use the canonical agent names in this prompt and fall back to current-agent execution if a specialist is unavailable. + +1. **Current State Assessment** — Invoke `SystemArchitect` via `#runSubagent` to document: - Current architecture topology and dependencies - Data stores, volumes, and access patterns - Integration points that will be affected @@ -25,14 +27,14 @@ Coordinate multi-agent migration planning: - `AzureStaticWebAppsSpecialist` — Frontend hosting migration (Vercel/Netlify → SWA, S3+CloudFront → SWA) 3. **Application Changes** — Invoke language specialists via `#runSubagent` to assess code impact: - - `python-specialist` / `rust-specialist` / `typescript-specialist` — SDK changes, connection string updates, feature parity gaps + - `PythonDeveloper` / `RustDeveloper` / `TypeScriptDeveloper` — SDK changes, connection string updates, feature parity gaps -4. **Infrastructure Changes** — Invoke `platform-quality` via `#runSubagent` to plan: +4. **Infrastructure Changes** — Invoke `PlatformEngineer` via `#runSubagent` to plan: - IaC templates (Bicep/Terraform) for the target architecture - CI/CD pipeline changes for new deployment targets - DNS cutover, certificate provisioning, and network routing -5. **Risk Analysis** — Invoke `risk-analysis-agent` via `#runSubagent` to evaluate: +5. **Risk Analysis** — Invoke `RiskAnalyst` via `#runSubagent` to evaluate: - Data loss risk during migration - Downtime windows and rollback procedures - Feature parity gaps between source and target diff --git a/.github/prompts/azure-troubleshoot.prompt.md b/.github/prompts/azure-troubleshoot.prompt.md index 1af9b0e..083d47b 100644 --- a/.github/prompts/azure-troubleshoot.prompt.md +++ b/.github/prompts/azure-troubleshoot.prompt.md @@ -7,6 +7,8 @@ input: "Describe the issue: which Azure service, error messages, affected endpoi Coordinate Azure service troubleshooting: +**Delegation compatibility**: If the workspace has `.github/agents/data/team-mapping.md`, resolve exact agent names from it before calling `#runSubagent`. Otherwise, use the canonical agent names in this prompt and fall back to current-agent execution if a specialist is unavailable. + 1. **Symptom Classification** — Identify the failing service type and route to the appropriate specialist via `#runSubagent`: - `AzureKubernetesSpecialist` — Pod crashes, image pull failures, node pressure, network policy blocks - `AzureAPIMSpecialist` — 502/504 errors, policy evaluation failures, backend timeouts, certificate issues @@ -18,7 +20,7 @@ Coordinate Azure service troubleshooting: - `AzureRedisSpecialist` — Cache misses, eviction storms, connection timeouts, memory pressure - `AzureStaticWebAppsSpecialist` — Build failures, routing mismatches, auth callback errors, API cold starts -2. **Log Analysis** — Invoke `platform-quality` via `#runSubagent` to: +2. **Log Analysis** — Invoke `PlatformEngineer` via `#runSubagent` to: - Query Azure Monitor / Log Analytics with KQL for the relevant error patterns - Check resource health and recent Azure status incidents - Review activity log for recent configuration changes that may have caused the regression @@ -30,11 +32,11 @@ Coordinate Azure service troubleshooting: - Code regression (bad deployment, missing env var) 4. **Fix Implementation** — Delegate the fix to the appropriate agent: - - Infrastructure fix → `platform-quality` - - Application code fix → `python-specialist` / `rust-specialist` / `typescript-specialist` + - Infrastructure fix → `PlatformEngineer` + - Application code fix → `PythonDeveloper` / `RustDeveloper` / `TypeScriptDeveloper` - Configuration fix → relevant Azure specialist -5. **Verification** — Invoke `platform-quality` to confirm: +5. **Verification** — Invoke `PlatformEngineer` to confirm: - Error rate returns to baseline - Health checks pass - No new alerts triggered diff --git a/.github/prompts/codebase-onboarding.prompt.md b/.github/prompts/codebase-onboarding.prompt.md index 1e4d817..c31ee67 100644 --- a/.github/prompts/codebase-onboarding.prompt.md +++ b/.github/prompts/codebase-onboarding.prompt.md @@ -7,7 +7,7 @@ input: "Optionally specify focus areas (backend, frontend, infra, a specific ser Coordinate a multi-agent codebase onboarding scan: -1. **Architecture Overview** — Invoke `system-architect` via `#runSubagent` to map: +1. **Architecture Overview** — Invoke `SystemArchitect` via `#runSubagent` to map: - High-level architecture (services, databases, external dependencies) - Communication patterns (sync HTTP, async events, shared DB) - Entry points (API endpoints, CLI commands, event handlers, scheduled jobs) @@ -15,17 +15,17 @@ Coordinate a multi-agent codebase onboarding scan: - Generate a Mermaid architecture diagram 2. **Technology Stack** — Invoke the relevant language specialist(s) via `#runSubagent`: - - `python-specialist` — Python version, framework (FastAPI/Django/Flask), package manager, key libraries - - `rust-specialist` — Rust edition, workspace structure, key crates, build targets - - `typescript-specialist` — Node/Deno/Bun runtime, framework (Next.js/Express/Remix), bundler, key packages + - `PythonDeveloper` — Python version, framework (FastAPI/Django/Flask), package manager, key libraries + - `RustDeveloper` — Rust edition, workspace structure, key crates, build targets + - `TypeScriptDeveloper` — Node/Deno/Bun runtime, framework (Next.js/Express/Remix), bundler, key packages -3. **Infrastructure & DevOps** — Invoke `platform-quality` via `#runSubagent` to document: +3. **Infrastructure & DevOps** — Invoke `PlatformEngineer` via `#runSubagent` to document: - How to run locally (setup steps, env vars, prerequisites) - How to run tests (commands, fixtures, test databases) - How to deploy (CI/CD pipeline, manual steps, environments) - IaC overview (Bicep/Terraform/CDK — what manages what) -4. **UI Layer** — If applicable, invoke `ui-agent` via `#runSubagent` to assess: +4. **UI Layer** — If applicable, invoke `UIDesigner` via `#runSubagent` to assess: - Component structure and design system - Styling approach (Tailwind, CSS modules, styled-components) - Accessibility baseline (initial WCAG read) @@ -42,3 +42,4 @@ Coordinate a multi-agent codebase onboarding scan: - Key files and directories to know - Common gotchas and pitfalls - Suggested first tasks for new contributors + diff --git a/.github/prompts/connector-audit.prompt.md b/.github/prompts/connector-audit.prompt.md index 59ada7d..86b8328 100644 --- a/.github/prompts/connector-audit.prompt.md +++ b/.github/prompts/connector-audit.prompt.md @@ -21,15 +21,15 @@ Coordinate an enterprise integration audit: - Dead-letter handling for failed messages 3. **Code Quality** — Invoke the relevant language specialist via `#runSubagent`: - - `python-specialist` / `rust-specialist` / `typescript-specialist` — Review connector code for error handling, type safety, test coverage + - `PythonDeveloper` / `RustDeveloper` / `TypeScriptDeveloper` — Review connector code for error handling, type safety, test coverage -4. **Security & Compliance** — Invoke `risk-analysis-agent` via `#runSubagent` to check: +4. **Security & Compliance** — Invoke `RiskAnalyst` via `#runSubagent` to check: - Credential rotation policy (are secrets expired or long-lived?) - Data handling compliance (PII masking, retention policies, GDPR/LGPD) - Audit trail completeness (can we trace every data exchange?) - Least-privilege access (are API scopes minimal?) -5. **Operational Health** — Invoke `platform-quality` via `#runSubagent` to verify: +5. **Operational Health** — Invoke `PlatformEngineer` via `#runSubagent` to verify: - Monitoring coverage (alerts on failure, latency, throughput drop) - Logging (structured, correlation IDs, no PII in logs) - Runbook exists for each integration failure scenario @@ -40,3 +40,4 @@ Coordinate an enterprise integration audit: - Critical findings requiring immediate remediation - Improvement recommendations ranked by risk reduction - Suggested SLA definitions for each integration + diff --git a/.github/prompts/connector-design.prompt.md b/.github/prompts/connector-design.prompt.md index 8a60a7e..d18ecce 100644 --- a/.github/prompts/connector-design.prompt.md +++ b/.github/prompts/connector-design.prompt.md @@ -13,19 +13,19 @@ Coordinate enterprise connector design: - Document data models and field mappings (source → internal → target) - Classify the platform category (PIM, DAM, CRM, Commerce, Inventory/SCM, Data/Analytics, Integration, Workforce, Identity, Privacy) -2. **Architecture Design** — Invoke `system-architect` via `#runSubagent` to: +2. **Architecture Design** — Invoke `SystemArchitect` via `#runSubagent` to: - Define the connector's place in the overall integration architecture - Choose communication pattern (sync request-response, async events, batch ETL, webhook-driven) - Design error handling strategy (retry policies, dead-letter queues, circuit breakers) - Specify contract versioning approach for the connector API surface 3. **Implementation Plan** — Invoke the appropriate language specialist via `#runSubagent`: - - `python-specialist` — For Python async connectors (httpx, aiohttp) - - `rust-specialist` — For high-throughput connectors (reqwest, tokio) - - `typescript-specialist` — For Node.js connectors (fetch, graphql-request) + - `PythonDeveloper` — For Python async connectors (httpx, aiohttp) + - `RustDeveloper` — For high-throughput connectors (reqwest, tokio) + - `TypeScriptDeveloper` — For Node.js connectors (fetch, graphql-request) - Define: auth flow, pagination handling, rate limit respect, data transformation pipeline -4. **Security Review** — Invoke `platform-quality` via `#runSubagent` to verify: +4. **Security Review** — Invoke `PlatformEngineer` via `#runSubagent` to verify: - Credential storage (Key Vault / env vars, never hardcoded) - OAuth token refresh handling - Input validation on all external data @@ -37,3 +37,4 @@ Coordinate enterprise connector design: - Load tests to validate rate limit handling Deliver a connector design document with architecture diagram, data mapping table, implementation spec, and test plan. + diff --git a/.github/prompts/migration-plan.prompt.md b/.github/prompts/migration-plan.prompt.md index 5c768de..c8aa4cc 100644 --- a/.github/prompts/migration-plan.prompt.md +++ b/.github/prompts/migration-plan.prompt.md @@ -7,30 +7,30 @@ input: "Describe the migration: source version/framework → target version/fram Coordinate a multi-agent migration plan: -1. **Impact Analysis** — Invoke `system-architect` via `#runSubagent` to assess: +1. **Impact Analysis** — Invoke `SystemArchitect` via `#runSubagent` to assess: - Which components are affected by the migration - Breaking changes between source and target (changelog review) - Integration points that may need contract updates - Dependency graph: what else must upgrade as a consequence 2. **Code Migration** — Invoke the relevant language specialist(s) via `#runSubagent`: - - `python-specialist` — Python version bumps, framework migrations (Django 4→5, FastAPI major), async runtime changes - - `rust-specialist` — Rust edition upgrades, major crate bumps, MSRV changes, cargo workspace restructures - - `typescript-specialist` — Node.js LTS upgrades, React major versions, Next.js pages→app router, bundler migrations + - `PythonDeveloper` — Python version bumps, framework migrations (Django 4→5, FastAPI major), async runtime changes + - `RustDeveloper` — Rust edition upgrades, major crate bumps, MSRV changes, cargo workspace restructures + - `TypeScriptDeveloper` — Node.js LTS upgrades, React major versions, Next.js pages→app router, bundler migrations - For each: identify deprecated APIs, removed features, new required patterns, and codemods if available -3. **Infrastructure Migration** — Invoke `platform-quality` via `#runSubagent` to plan: +3. **Infrastructure Migration** — Invoke `PlatformEngineer` via `#runSubagent` to plan: - CI/CD pipeline changes (new tool versions, updated actions, changed test commands) - IaC updates (new provider versions, resource API changes) - Container base image updates (new runtime version) - Environment variable or configuration format changes -4. **UI Migration** — If applicable, invoke `ui-agent` via `#runSubagent`: +4. **UI Migration** — If applicable, invoke `UIDesigner` via `#runSubagent`: - Component library upgrade (breaking style changes, removed components) - CSS framework migration (Tailwind v3→v4, CSS-in-JS → Tailwind) - Accessibility regression check after visual changes -5. **Risk Assessment** — Invoke `risk-analysis-agent` via `#runSubagent`: +5. **Risk Assessment** — Invoke `RiskAnalyst` via `#runSubagent`: - Categorize risks: data loss, downtime, behavioral change, performance regression - Identify rollback strategy for each migration stage - Flag migration steps that are irreversible @@ -43,3 +43,4 @@ Coordinate a multi-agent migration plan: - **Stage 5: Cleanup** — Remove compatibility shims, deprecated code paths, and old config - Rollback trigger criteria and procedure for each stage - Estimated timeline per stage + diff --git a/.github/prompts/platform-cicd-audit.prompt.md b/.github/prompts/platform-cicd-audit.prompt.md index e35ac98..2e6db8d 100644 --- a/.github/prompts/platform-cicd-audit.prompt.md +++ b/.github/prompts/platform-cicd-audit.prompt.md @@ -1,7 +1,7 @@ --- title: "Platform: CI/CD Audit" description: "Audit CI/CD pipelines for reliability, security, test masking, and deployment safety." -mode: "platform-quality" +mode: "PlatformEngineer" input: "Specify the pipeline file(s) or CI system (GitHub Actions, Azure Pipelines) to audit." --- diff --git a/.github/prompts/platform-dependency-audit.prompt.md b/.github/prompts/platform-dependency-audit.prompt.md index e6eacbc..20c6903 100644 --- a/.github/prompts/platform-dependency-audit.prompt.md +++ b/.github/prompts/platform-dependency-audit.prompt.md @@ -1,7 +1,7 @@ --- title: "Platform: Dependency Audit" description: "Audit project dependencies for vulnerabilities, currency, and license compliance." -mode: "platform-quality" +mode: "PlatformEngineer" input: "Specify the project or lockfile to audit (package.json, Cargo.toml, pyproject.toml, etc.)." --- diff --git a/.github/prompts/platform-quality-evaluation.prompt.md b/.github/prompts/platform-quality-evaluation.prompt.md index fd2c07e..84ef68a 100644 --- a/.github/prompts/platform-quality-evaluation.prompt.md +++ b/.github/prompts/platform-quality-evaluation.prompt.md @@ -7,24 +7,24 @@ input: "Specify the project or repository to evaluate. Optionally narrow scope t Conduct a full platform quality evaluation by delegating to specialist agents: -1. **CI/CD & Infrastructure** — Invoke `platform-quality` via `#runSubagent` to audit: +1. **CI/CD & Infrastructure** — Invoke `PlatformEngineer` via `#runSubagent` to audit: - Pipeline reliability (test masking, error handling, timeout configuration) - Security posture (secrets management, SAST integration, least-privilege tokens) - Dependency hygiene (vulnerability scan, version currency, license compliance) - Environment promotion strategy and deployment safety -2. **Architecture Compliance** — Invoke `system-architect` via `#runSubagent` to evaluate: +2. **Architecture Compliance** — Invoke `SystemArchitect` via `#runSubagent` to evaluate: - ADR adherence (are documented decisions being followed?) - Dependency direction (domain layer has no infra imports?) - Integration patterns (event contracts versioned? circuit breakers present?) - Component boundaries (single responsibility, no god services) 3. **Code Quality by Language** — Invoke the appropriate specialist per language stack: - - `python-specialist` — Type coverage, async correctness, test quality, PEP compliance - - `rust-specialist` — Unsafe audit, error handling patterns, clippy compliance, cargo audit - - `typescript-specialist` — Strict mode compliance, ESLint zero-warnings, React patterns, bundle analysis + - `PythonDeveloper` — Type coverage, async correctness, test quality, PEP compliance + - `RustDeveloper` — Unsafe audit, error handling patterns, clippy compliance, cargo audit + - `TypeScriptDeveloper` — Strict mode compliance, ESLint zero-warnings, React patterns, bundle analysis -4. **Accessibility & UI** — Invoke `ui-agent` via `#runSubagent` to audit: +4. **Accessibility & UI** — Invoke `UIDesigner` via `#runSubagent` to audit: - WCAG 2.2 AA compliance across UI surfaces - Semantic HTML structure and keyboard navigation - Core Web Vitals performance (LCP, INP, CLS) @@ -35,3 +35,4 @@ Conduct a full platform quality evaluation by delegating to specialist agents: - Comparison with previous evaluation if available Deliver a quality report with per-category scores, top 10 remediation items, and trend indicators. + diff --git a/.github/prompts/pr-review.prompt.md b/.github/prompts/pr-review.prompt.md index 7e643ea..0d4496f 100644 --- a/.github/prompts/pr-review.prompt.md +++ b/.github/prompts/pr-review.prompt.md @@ -1,7 +1,7 @@ --- title: "PR: Review Pull Request" description: "Review a pull request for architecture compliance, test coverage, security, and merge readiness." -mode: "pr-evaluator" +mode: "PRReviewer" input: "Provide the PR number or let the agent detect the active PR. Optionally specify focus areas." --- diff --git a/.github/prompts/python-implement.prompt.md b/.github/prompts/python-implement.prompt.md index 15a0e47..9cc2230 100644 --- a/.github/prompts/python-implement.prompt.md +++ b/.github/prompts/python-implement.prompt.md @@ -1,7 +1,7 @@ --- title: "Python: Implement Feature" description: "Implement a Python feature with type hints, async patterns, and pytest coverage following PEP standards." -mode: "python-specialist" +mode: "PythonDeveloper" input: "Describe the feature, module, or function to implement. Include any API contracts or data models." --- diff --git a/.github/prompts/python-refactor.prompt.md b/.github/prompts/python-refactor.prompt.md index 5b0d0fc..e931a91 100644 --- a/.github/prompts/python-refactor.prompt.md +++ b/.github/prompts/python-refactor.prompt.md @@ -1,7 +1,7 @@ --- title: "Python: Refactor Module" description: "Refactor Python code to reduce complexity, eliminate code smells, and improve maintainability." -mode: "python-specialist" +mode: "PythonDeveloper" input: "Specify the module or file to refactor. Optionally describe the target improvement (performance, readability, testability)." --- diff --git a/.github/prompts/python-review.prompt.md b/.github/prompts/python-review.prompt.md index e8975ff..0f34bea 100644 --- a/.github/prompts/python-review.prompt.md +++ b/.github/prompts/python-review.prompt.md @@ -1,7 +1,7 @@ --- title: "Python: Review Code" description: "Review Python code for type safety, async correctness, security, and adherence to PEP standards." -mode: "python-specialist" +mode: "PythonDeveloper" input: "Provide the file(s) or module to review. Optionally specify focus areas (security, performance, correctness)." --- diff --git a/.github/prompts/rust-implement.prompt.md b/.github/prompts/rust-implement.prompt.md index 0ad89b7..63f69fa 100644 --- a/.github/prompts/rust-implement.prompt.md +++ b/.github/prompts/rust-implement.prompt.md @@ -1,7 +1,7 @@ --- title: "Rust: Implement Feature" description: "Implement a Rust feature with safe ownership patterns, proper error handling, and cargo test coverage." -mode: "rust-specialist" +mode: "RustDeveloper" input: "Describe the feature, crate, or module to implement. Include any trait contracts or wire formats." --- diff --git a/.github/prompts/rust-review.prompt.md b/.github/prompts/rust-review.prompt.md index a41f7e0..0632fb3 100644 --- a/.github/prompts/rust-review.prompt.md +++ b/.github/prompts/rust-review.prompt.md @@ -1,7 +1,7 @@ --- title: "Rust: Review Code" description: "Review Rust code for safety, ownership correctness, performance, and idiomatic patterns." -mode: "rust-specialist" +mode: "RustDeveloper" input: "Provide the crate or file(s) to review. Optionally specify focus (safety, performance, API design)." --- diff --git a/.github/prompts/security-audit.prompt.md b/.github/prompts/security-audit.prompt.md index 90b6b1d..cfbde3f 100644 --- a/.github/prompts/security-audit.prompt.md +++ b/.github/prompts/security-audit.prompt.md @@ -7,18 +7,18 @@ input: "Specify the scope: entire repo, specific service, or specific concern (O Coordinate a comprehensive security audit by delegating to all relevant specialists: -1. **Dependency Vulnerabilities** — Invoke `platform-quality` via `#runSubagent` to: +1. **Dependency Vulnerabilities** — Invoke `PlatformEngineer` via `#runSubagent` to: - Run all package audits (npm audit, cargo audit, pip-audit, trivy) - Classify findings by severity (critical/high/medium/low) - Identify transitive dependency risks (nested vulnerabilities) - Check for abandoned packages (no updates in 12+ months) 2. **Application Code** — Invoke language specialists via `#runSubagent` to audit OWASP Top 10: - - `python-specialist` — SQL injection (parameterized queries), SSRF, insecure deserialization, auth bypass - - `rust-specialist` — Unsafe blocks, FFI boundary validation, panic paths in production, input parsing - - `typescript-specialist` — XSS (dangerouslySetInnerHTML), CSRF, prototype pollution, JWT validation + - `PythonDeveloper` — SQL injection (parameterized queries), SSRF, insecure deserialization, auth bypass + - `RustDeveloper` — Unsafe blocks, FFI boundary validation, panic paths in production, input parsing + - `TypeScriptDeveloper` — XSS (dangerouslySetInnerHTML), CSRF, prototype pollution, JWT validation -3. **Infrastructure & Secrets** — Invoke `platform-quality` via `#runSubagent` to check: +3. **Infrastructure & Secrets** — Invoke `PlatformEngineer` via `#runSubagent` to check: - No secrets in code, env files, or CI logs (scan for patterns: API keys, tokens, connection strings) - Secret rotation policy (Key Vault, GitHub Secrets — age and expiry) - CI/CD pipeline security (least-privilege tokens, no --no-verify, signed commits) @@ -30,7 +30,7 @@ Coordinate a comprehensive security audit by delegating to all relevant speciali - Encryption (at rest, in transit, customer-managed keys if required) - Logging (diagnostic settings, audit trails, retention) -5. **Architecture** — Invoke `system-architect` via `#runSubagent` to review: +5. **Architecture** — Invoke `SystemArchitect` via `#runSubagent` to review: - Trust boundaries (where does the system trust external input?) - Authentication flow (token lifecycle, session management) - Authorization model (RBAC, ABAC — is least privilege enforced?) @@ -42,3 +42,4 @@ Coordinate a comprehensive security audit by delegating to all relevant speciali - Critical/high findings requiring immediate action - Compliance gap analysis (against specified standards) - Remediation backlog with effort estimates + diff --git a/.github/prompts/ship-deploy-hotfix.prompt.md b/.github/prompts/ship-deploy-hotfix.prompt.md index 718e6db..00561c0 100644 --- a/.github/prompts/ship-deploy-hotfix.prompt.md +++ b/.github/prompts/ship-deploy-hotfix.prompt.md @@ -19,24 +19,24 @@ Delegate implementation and validation to the appropriate specialist agents via | Change Type | Agent | Validation | |---|---|---| -| Python code | `python-specialist` | Type check (mypy), tests (pytest), lint | -| Rust code | `rust-specialist` | cargo check, cargo clippy, cargo test | -| TypeScript/React | `typescript-specialist` | tsc --noEmit, ESLint, Vitest | -| UI/accessibility | `ui-agent` | WCAG 2.2 AA audit, visual regression | -| CI/CD / infra | `platform-quality` | actionlint, dry-run deploy, IaC validate | -| Architecture impact | `system-architect` | ADR compliance, integration contract review | +| Python code | `PythonDeveloper` | Type check (mypy), tests (pytest), lint | +| Rust code | `RustDeveloper` | cargo check, cargo clippy, cargo test | +| TypeScript/React | `TypeScriptDeveloper` | tsc --noEmit, ESLint, Vitest | +| UI/accessibility | `UIDesigner` | WCAG 2.2 AA audit, visual regression | +| CI/CD / infra | `PlatformEngineer` | actionlint, dry-run deploy, IaC validate | +| Architecture impact | `SystemArchitect` | ADR compliance, integration contract review | | Azure resources | Relevant Azure specialist | Resource validation, config drift check | Run **all** applicable validations — most changes touch multiple domains. ## Phase 3: Pull Request -1. **Create PR** — Invoke `pr-evaluator` via `#runSubagent` to: +1. **Create PR** — Invoke `PRReviewer` via `#runSubagent` to: - Open the PR referencing the issue (`Closes #`) - Verify all CI checks pass - Confirm test coverage is non-negative - Validate no security regressions (OWASP Top 10) - - Obtain architecture sign-off from `system-architect` if the change crosses service boundaries + - Obtain architecture sign-off from `SystemArchitect` if the change crosses service boundaries 2. **Review Gate** — PR must have: - [ ] All checks green - [ ] No unresolved review comments @@ -45,7 +45,7 @@ Run **all** applicable validations — most changes touch multiple domains. ## Phase 4: Merge & Deploy 1. **Merge** — Squash-merge into the default branch. Ensure commit message references the issue. -2. **Deploy** — Invoke `platform-quality` via `#runSubagent` to: +2. **Deploy** — Invoke `PlatformEngineer` via `#runSubagent` to: - Trigger the deployment pipeline - Monitor deployment progress (build → staging → production) - Verify health checks pass post-deploy @@ -54,7 +54,7 @@ Run **all** applicable validations — most changes touch multiple domains. After deployment, actively monitor for regressions: -1. **Health Check** — Invoke `platform-quality` to verify: +1. **Health Check** — Invoke `PlatformEngineer` to verify: - Application health endpoints return 200 - No new errors in logs (compare error rate before/after) - Performance metrics within baseline (latency, throughput) @@ -65,7 +65,7 @@ After deployment, actively monitor for regressions: - **P1/P2: Immediate hotfix** — - Create a `hotfix/-` branch from the default branch - Invoke the responsible specialist agent to implement the fix - - Fast-track PR: `pr-evaluator` reviews with expedited checklist (tests pass + no regression) + - Fast-track PR: `PRReviewer` reviews with expedited checklist (tests pass + no regression) - Merge and redeploy immediately - **P3/P4: Deferred fix** — - Create a new issue with `[Hotfix]` prefix @@ -76,7 +76,7 @@ After deployment, actively monitor for regressions: - Root cause and causal chain - Why existing tests didn't catch it - Prevention measures (new tests, monitoring, validation steps) - - Delegate prevention implementation to `platform-quality` + - Delegate prevention implementation to `PlatformEngineer` ## Completion Criteria @@ -86,3 +86,4 @@ After deployment, actively monitor for regressions: - [ ] Deployment successful with health checks passing - [ ] No P1/P2 regressions within monitoring window - [ ] Any hotfixes completed and documented + diff --git a/.github/prompts/tech-debt-assessment.prompt.md b/.github/prompts/tech-debt-assessment.prompt.md index 6cd2190..abb96d8 100644 --- a/.github/prompts/tech-debt-assessment.prompt.md +++ b/.github/prompts/tech-debt-assessment.prompt.md @@ -8,25 +8,25 @@ input: "Specify scope: entire repo or specific areas. Optionally include known p Coordinate a systematic tech debt inventory: 1. **Code-Level Debt** — Invoke language specialists via `#runSubagent` to scan for: - - `python-specialist` — Missing type hints, bare except blocks, deprecated APIs, TODO/FIXME/HACK comments, untested code paths - - `rust-specialist` — Clippy warnings suppressed, unwrap() in non-test code, outdated edition, unsafe without justification - - `typescript-specialist` — `any` type usage, disabled ESLint rules, legacy class components, untyped APIs, outdated dependencies + - `PythonDeveloper` — Missing type hints, bare except blocks, deprecated APIs, TODO/FIXME/HACK comments, untested code paths + - `RustDeveloper` — Clippy warnings suppressed, unwrap() in non-test code, outdated edition, unsafe without justification + - `TypeScriptDeveloper` — `any` type usage, disabled ESLint rules, legacy class components, untyped APIs, outdated dependencies -2. **Architecture Debt** — Invoke `system-architect` via `#runSubagent` to identify: +2. **Architecture Debt** — Invoke `SystemArchitect` via `#runSubagent` to identify: - Circular dependencies between modules/services - God classes/modules with too many responsibilities - Missing abstraction layers (business logic in controllers, SQL in handlers) - Stale ADRs that no longer reflect reality - Integration contracts that have drifted from specification -3. **Infrastructure Debt** — Invoke `platform-quality` via `#runSubagent` to catalog: +3. **Infrastructure Debt** — Invoke `PlatformEngineer` via `#runSubagent` to catalog: - Manual deployment steps that should be automated - Missing or flaky tests in CI/CD pipeline - Outdated base images, deprecated GitHub Actions, legacy IaC patterns - Monitoring gaps (services without health checks, missing alerts) - Documentation that's out of date with actual behavior -4. **UI Debt** — If applicable, invoke `ui-agent` via `#runSubagent` to identify: +4. **UI Debt** — If applicable, invoke `UIDesigner` via `#runSubagent` to identify: - Accessibility violations accumulated over time - Inconsistent component patterns (multiple button styles, duplicated layouts) - Performance regressions (large bundles, layout shifts, slow interactions) @@ -44,3 +44,4 @@ Coordinate a systematic tech debt inventory: - Quick wins (< 1 day effort, high impact) highlighted separately - Estimated total remediation effort (in engineering days) - Recommended cadence for debt reduction (e.g., 20% of each sprint) + diff --git a/.github/prompts/tech-lead-document-architecture.prompt.md b/.github/prompts/tech-lead-document-architecture.prompt.md index e4d5c52..369d2ad 100644 --- a/.github/prompts/tech-lead-document-architecture.prompt.md +++ b/.github/prompts/tech-lead-document-architecture.prompt.md @@ -7,33 +7,34 @@ input: "Specify the system or project to document. Include target audience (deve Coordinate multi-agent documentation of the platform: -1. **Business Context** — Invoke `business-strategy-agent` via `#runSubagent` to document: +1. **Business Context** — Invoke `BusinessStrategist` via `#runSubagent` to document: - System purpose and business value proposition - Key stakeholders and their concerns - Domain boundaries and bounded contexts - Non-functional requirements (SLAs, compliance, scaling targets) -2. **Architecture Documentation** — Invoke `system-architect` via `#runSubagent` to produce: +2. **Architecture Documentation** — Invoke `SystemArchitect` via `#runSubagent` to produce: - Architecture Decision Records (ADRs) for every non-trivial decision - Component responsibility matrix - Data flow and integration contracts between services - Technology stack rationale and trade-off analysis -3. **Visual Architecture** — Invoke `ui-agent` via `#runSubagent` to create: +3. **Visual Architecture** — Invoke `UIDesigner` via `#runSubagent` to create: - **C4 diagrams** (Context, Container, Component, Code) as draw.io XML files - **High-level architecture** overview as a Mermaid diagram with consistent theming - **Sequence diagrams** for critical flows (auth, data pipeline, deployment) - **Design principia**: icon set selection, color palette definition, diagram style guide ensuring visual consistency across all artifacts 4. **Technical Reference** — Invoke language specialists via `#runSubagent` as needed: - - `python-specialist` — API reference, async patterns guide, data model docs - - `rust-specialist` — Crate architecture, safety invariants, performance contracts - - `typescript-specialist` — Component catalog, state management patterns, route structure + - `PythonDeveloper` — API reference, async patterns guide, data model docs + - `RustDeveloper` — Crate architecture, safety invariants, performance contracts + - `TypeScriptDeveloper` — Component catalog, state management patterns, route structure -5. **Operations Runbook** — Invoke `platform-quality` via `#runSubagent` to document: +5. **Operations Runbook** — Invoke `PlatformEngineer` via `#runSubagent` to document: - Deployment procedures and rollback steps - Monitoring and alerting configuration - Incident response playbook - Environment setup and prerequisites Deliver a documentation package with ADRs, diagrams (Mermaid + draw.io), API reference, and runbook organized by audience. + diff --git a/.github/prompts/tech-lead-innovation-research.prompt.md b/.github/prompts/tech-lead-innovation-research.prompt.md index a43af60..8cfb808 100644 --- a/.github/prompts/tech-lead-innovation-research.prompt.md +++ b/.github/prompts/tech-lead-innovation-research.prompt.md @@ -7,21 +7,23 @@ input: "Describe the problem space or opportunity to explore. Include business c Coordinate a multi-agent innovation research cycle: +**Delegation compatibility**: If the workspace has `.github/agents/data/team-mapping.md`, resolve exact agent names from it before calling `#runSubagent`. Otherwise, use the canonical agent names in this prompt and fall back to current-agent execution if a specialist is unavailable. + 1. **Opportunity Framing** — Invoke business agents via `#runSubagent` to define the innovation space: - - `business-strategy-agent` — Market positioning, competitive landscape, strategic fit - - `competitive-intelligence-agent` — What competitors and adjacent industries are doing differently - - `financial-modeling-agent` — Cost-benefit modeling, ROI projections for candidate solutions - - `risk-analysis-agent` — Risk assessment for each candidate (technical, market, regulatory) + - `BusinessStrategist` — Market positioning, competitive landscape, strategic fit + - `CompetitiveIntelAnalyst` — What competitors and adjacent industries are doing differently + - `FinancialModeler` — Cost-benefit modeling, ROI projections for candidate solutions + - `RiskAnalyst` — Risk assessment for each candidate (technical, market, regulatory) -2. **State of the Art** — Invoke `Explore` via `#runSubagent` to survey: +2. **State of the Art** — Use the available research, workspace, and web tools directly to survey: - Recent academic papers (arXiv, ACM, IEEE) for novel approaches in the problem space - Community adoption trends, emerging frameworks, open-source momentum - Comparable implementations in open-source projects 3. **Technical Feasibility** — Invoke engineering agents via `#runSubagent` to evaluate: - - `system-architect` — Architectural fit, integration complexity, migration path from current state - - `python-specialist` / `rust-specialist` / `typescript-specialist` — Prototype feasibility, language ecosystem maturity, library availability - - `platform-quality` — Infrastructure requirements, deployment complexity, operational burden + - `SystemArchitect` — Architectural fit, integration complexity, migration path from current state + - `PythonDeveloper` / `RustDeveloper` / `TypeScriptDeveloper` — Prototype feasibility, language ecosystem maturity, library availability + - `PlatformEngineer` — Infrastructure requirements, deployment complexity, operational burden 4. **Synthesis** — Combine business and technical findings into a decision matrix: - Score each candidate on: business value, technical feasibility, time-to-value, operational cost, risk diff --git a/.github/prompts/tech-lead-investigate.prompt.md b/.github/prompts/tech-lead-investigate.prompt.md index b197a22..9737c3e 100644 --- a/.github/prompts/tech-lead-investigate.prompt.md +++ b/.github/prompts/tech-lead-investigate.prompt.md @@ -10,13 +10,14 @@ Investigate the reported issue: 1. **Symptom Analysis** — Restate the problem. Identify affected components and user impact severity. 2. **Hypothesis Formation** — List the 3 most likely root causes ranked by probability. 3. **Evidence Gathering** — For each hypothesis, invoke the appropriate specialist agent via `#runSubagent` to gather evidence: - - `python-specialist` — for Python stack traces, async bugs, type errors - - `rust-specialist` — for Rust panics, ownership issues, linker failures - - `typescript-specialist` — for TypeScript type errors, React rendering, bundle issues - - `platform-quality` — for CI/CD failures, infrastructure, dependency problems - - `system-architect` — for cross-service integration failures, architectural mismatches + - `PythonDeveloper` — for Python stack traces, async bugs, type errors + - `RustDeveloper` — for Rust panics, ownership issues, linker failures + - `TypeScriptDeveloper` — for TypeScript type errors, React rendering, bundle issues + - `PlatformEngineer` — for CI/CD failures, infrastructure, dependency problems + - `SystemArchitect` — for cross-service integration failures, architectural mismatches 4. **Root Cause Isolation** — Narrow to the confirmed cause. Document the causal chain. 5. **Fix Planning** — Decompose the fix into agent-assignable tasks. Invoke each specialist via `#runSubagent` to implement their portion. -6. **Prevention** — Delegate to `platform-quality` to recommend test/monitoring additions to prevent recurrence. +6. **Prevention** — Delegate to `PlatformEngineer` to recommend test/monitoring additions to prevent recurrence. Deliver an investigation report with root cause, fix plan, and prevention recommendations. + diff --git a/.github/prompts/tech-lead-issue-correction.prompt.md b/.github/prompts/tech-lead-issue-correction.prompt.md index d0d969f..eb62acb 100644 --- a/.github/prompts/tech-lead-issue-correction.prompt.md +++ b/.github/prompts/tech-lead-issue-correction.prompt.md @@ -15,13 +15,14 @@ Analyze the reported problems and create a structured remediation plan: - Acceptance criteria as a checklist - Labels for priority, component, and issue type 4. **Agent Assignment** — Assign each issue to the appropriate specialist agent via `#runSubagent`: - - `python-specialist` — Python bugs, type errors, async issues, test failures - - `rust-specialist` — Rust panics, ownership bugs, linker issues, unsafe violations - - `typescript-specialist` — TypeScript errors, React rendering bugs, bundle failures - - `ui-agent` — Accessibility violations, layout regressions, responsive breakage - - `platform-quality` — CI/CD failures, dependency vulnerabilities, infra misconfigurations - - `system-architect` — Integration contract violations, architectural drift + - `PythonDeveloper` — Python bugs, type errors, async issues, test failures + - `RustDeveloper` — Rust panics, ownership bugs, linker issues, unsafe violations + - `TypeScriptDeveloper` — TypeScript errors, React rendering bugs, bundle failures + - `UIDesigner` — Accessibility violations, layout regressions, responsive breakage + - `PlatformEngineer` — CI/CD failures, dependency vulnerabilities, infra misconfigurations + - `SystemArchitect` — Integration contract violations, architectural drift 5. **Dependency Sequencing** — Order issue resolution by dependency graph. Identify which fixes unblock others. 6. **Execution** — Invoke each specialist via `#runSubagent` with the issue brief. Track completion against the issue checklist. Deliver a table mapping each issue number to its agent, priority, dependencies, and status. + diff --git a/.github/prompts/tech-lead-issue-execute.prompt.md b/.github/prompts/tech-lead-issue-execute.prompt.md index beaaa43..e76dcfc 100644 --- a/.github/prompts/tech-lead-issue-execute.prompt.md +++ b/.github/prompts/tech-lead-issue-execute.prompt.md @@ -13,7 +13,7 @@ Drive a GitHub issue from open to closed: - Scope boundary: what is explicitly in and out of scope - Priority and severity: infer from labels, description, or ask -2. **Codebase Reconnaissance** — Invoke `Explore` via `#runSubagent` to: +2. **Codebase Reconnaissance** — Use workspace search and file-reading tools directly to: - Identify the files, modules, and services affected - Map dependencies and potential blast radius - Surface existing tests and related code @@ -23,27 +23,26 @@ Drive a GitHub issue from open to closed: - Identify dependencies between sub-tasks and sequence them - Estimate complexity (trivial / small / medium / large) -4. **Architecture Check** — If the issue crosses service boundaries or changes data flow, invoke `system-architect` via `#runSubagent` to: +4. **Architecture Check** — If the issue crosses service boundaries or changes data flow, invoke `SystemArchitect` via `#runSubagent` to: - Validate the approach against established patterns - Identify integration contracts that need updating - Flag any ADR that should be created or amended 5. **Implementation** — Delegate each sub-task to the appropriate specialist via `#runSubagent`: - - `python-specialist` — Python implementation, tests, type checking - - `rust-specialist` — Rust implementation, ownership correctness, cargo tests - - `typescript-specialist` — TypeScript/React implementation, type safety, Vitest - - `ui-agent` — UI components, accessibility, responsive design - - `platform-quality` — CI/CD changes, infrastructure, dependency updates + - `PythonDeveloper` — Python implementation, tests, type checking + - `RustDeveloper` — Rust implementation, ownership correctness, cargo tests + - `TypeScriptDeveloper` — TypeScript/React implementation, type safety, Vitest + - `UIDesigner` — UI components, accessibility, responsive design + - `PlatformEngineer` — CI/CD changes, infrastructure, dependency updates - `enterprise-connectors` — External API integrations, connector changes 6. **Verification** — After implementation: - Confirm all acceptance criteria are met - Run tests and verify no regressions - - Invoke `pr-evaluator` via `#runSubagent` for merge-readiness review + - Invoke `PRReviewer` via `#runSubagent` for merge-readiness review 7. **Closure** — Summarize what was done: - Changes made (files, modules, tests) - Decisions taken and rationale - Any follow-up items or deferred scope - -Deliver a completion report linking the issue to the changes, with verification evidence. + - Deliver a completion report linking the issue to the changes, with verification evidence. diff --git a/.github/prompts/tech-lead-plan.prompt.md b/.github/prompts/tech-lead-plan.prompt.md index 4c7383e..4e38e80 100644 --- a/.github/prompts/tech-lead-plan.prompt.md +++ b/.github/prompts/tech-lead-plan.prompt.md @@ -8,18 +8,19 @@ input: "Describe the feature or initiative. Include business context, scope boun Plan the requested feature: 1. **Business Context** — Restate the requirement. Connect to user outcome or business metric. -2. **Architecture Assessment** — Delegate to `system-architect` via `#runSubagent` to evaluate system impact, affected components, and integration points. +2. **Architecture Assessment** — Delegate to `SystemArchitect` via `#runSubagent` to evaluate system impact, affected components, and integration points. 3. **Task Decomposition** — Break into atomic, testable sub-tasks. Each must be completable by a single specialist agent. 4. **Agent Assignment** — Map each sub-task to the appropriate specialist agent. Use these exact agent names for delegation via `#runSubagent`: - - `python-specialist` — Python implementation - - `rust-specialist` — Rust implementation - - `typescript-specialist` — TypeScript/React implementation - - `ui-agent` — UI/UX design and accessibility - - `platform-quality` — CI/CD, IaC, and cross-cutting infrastructure - - `system-architect` — Architecture decisions, ADRs, pattern validation - - `pr-evaluator` — Final PR review before merge + - `PythonDeveloper` — Python implementation + - `RustDeveloper` — Rust implementation + - `TypeScriptDeveloper` — TypeScript/React implementation + - `UIDesigner` — UI/UX design and accessibility + - `PlatformEngineer` — CI/CD, IaC, and cross-cutting infrastructure + - `SystemArchitect` — Architecture decisions, ADRs, pattern validation + - `PRReviewer` — Final PR review before merge 5. **Dependency Graph** — Sequence tasks. Identify which can run in parallel vs which block others. 6. **Risk Assessment** — Document risks (breaking changes, security, performance, scope creep) with mitigations. 7. **Acceptance Criteria** — Define "done" for each sub-task and for the feature as a whole. Deliver a plan as a structured table with task ID, title, agent name, dependencies, acceptance criteria, and risk level. Then execute the plan by invoking each agent via `#runSubagent` in dependency order. + diff --git a/.github/prompts/typescript-implement.prompt.md b/.github/prompts/typescript-implement.prompt.md index 0f02d9c..42791f6 100644 --- a/.github/prompts/typescript-implement.prompt.md +++ b/.github/prompts/typescript-implement.prompt.md @@ -1,7 +1,7 @@ --- title: "TypeScript: Implement Feature" description: "Implement a TypeScript/React feature with strict types, React Server Components, and accessibility compliance." -mode: "typescript-specialist" +mode: "TypeScriptDeveloper" input: "Describe the feature or component to implement. Include API contracts, UI wireframe description, or data requirements." --- diff --git a/.github/prompts/typescript-review.prompt.md b/.github/prompts/typescript-review.prompt.md index 65561da..6ea3ee2 100644 --- a/.github/prompts/typescript-review.prompt.md +++ b/.github/prompts/typescript-review.prompt.md @@ -1,7 +1,7 @@ --- title: "TypeScript: Review Code" description: "Review TypeScript/React code for type safety, accessibility, performance, and React best practices." -mode: "typescript-specialist" +mode: "TypeScriptDeveloper" input: "Provide the file(s) or component tree to review. Optionally specify focus (types, a11y, performance)." --- diff --git a/.github/prompts/ui-accessibility-audit.prompt.md b/.github/prompts/ui-accessibility-audit.prompt.md index e2f184b..d0a4403 100644 --- a/.github/prompts/ui-accessibility-audit.prompt.md +++ b/.github/prompts/ui-accessibility-audit.prompt.md @@ -1,7 +1,7 @@ --- title: "UI: Accessibility Audit" description: "Audit a UI component or page for WCAG 2.2 AA compliance, keyboard navigation, and screen reader compatibility." -mode: "ui-agent" +mode: "UIDesigner" input: "Specify the component, page, or URL to audit. Include the target platform (web, CLI, desktop)." --- diff --git a/.github/prompts/ui-design-component.prompt.md b/.github/prompts/ui-design-component.prompt.md index 4180c8d..9573e25 100644 --- a/.github/prompts/ui-design-component.prompt.md +++ b/.github/prompts/ui-design-component.prompt.md @@ -1,7 +1,7 @@ --- title: "UI: Design Component" description: "Design and implement a responsive, accessible UI component with proper semantic HTML and design tokens." -mode: "ui-agent" +mode: "UIDesigner" input: "Describe the component (purpose, content, interactions). Specify platform: web (Tailwind), CLI (Ink), or desktop (Tauri)." ---