[rig-claude] Improve Claude dynamic-workflow compatibility for rig - #300
Merged
Merged
Conversation
…n table The primitive mapping table in claude-workflow-conversion.md was missing the until() convergence primitive, which is the idiomatic rig replacement for open-ended while loops in Claude dynamic workflows. Without this entry, developers porting a workflow that uses a bounded retry/convergence pattern would not discover until() and might write an unbounded loop instead. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
pelikhan
marked this pull request as ready for review
July 30, 2026 10:40
Contributor
Author
|
🧠 Matt Pocock Skills Reviewer has completed the skills-based review. ✅ |
Contributor
Author
There was a problem hiding this comment.
Skills-Based Review 🧠
Applied /grill-with-docs — this is a clean, accurate documentation addition.
📋 Key Themes & Highlights
Positive Highlights
- ✅ The new row correctly identifies
until({ max, noProgressRounds }, step)as the rig idiom for bounded convergence loops - ✅ Placement is consistent — inserted between the
workflow()row and top-level return, which is the logical location for a loop primitive - ✅ The notes column accurately summarises the three stopping conditions (
done,max,noProgressRounds) matchingdynamic-workflows.md - ✅ The "prefer over unbounded loops" guidance aligns with how
untilis presented in the reference docs
No issues found. The addition closes the last visible gap in the primitive mapping table.
🧠 Reviewed using Matt Pocock's skills by Matt Pocock Skills Reviewer · sonnet46 21.9 AIC · ⌖ 4.72 AIC · ⊞ 6.3K
Comment /matt to run again
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Compatibility gap addressed
The primitive mapping table in
skills/rig/references/claude-workflow-conversion.mdwas missing theuntil()convergence primitive. A developer porting a Claude dynamic workflow that uses an open-endedwhileloop for bounded retries or convergence (e.g. "keep fixing until tests pass, up to 8 rounds") would not discoveruntil()and might write an unbounded loop instead.Why this improves transfer
untilis the exact rig idiom for the "bounded convergence loop" pattern that is common in Claude dynamic workflows. The table already covered every other major primitive (parallel,pipeline,phase,log,budget, nestedworkflow, schema-constrained calls, per-call options). Addinguntilcloses the last visible gap in the conversion surface.Files changed
skills/rig/references/claude-workflow-conversion.md— added one row to the primitive mapping table:| open-ended
whileconvergence loop |until({ max, noProgressRounds }, step)| Bounded; stops ondone, aftermaxrounds, or afternoProgressRoundsequal progress keys — prefer over unbounded loops |Validation
Only documentation changed. Verified:
dynamic-workflows.md,agent-api.md, andruntime.mdunchanged).wc -lwent from 155 → 156 lines (one added row).Remaining intentional differences
All previously documented differences are unchanged:
{ effort: "high" }has no rig equivalent (encode structurally or in model id){ agentType: "Explore" }has no built-in read-only agent type