Loop review patch retest
Use to run a bounded review, patch, and retest loop that stops on green, on no diff, or on a repeated identical finding.From its SKILL.md
npx -y skills add jsuvic/agent-profile --skill loop-review-patch-retestAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.6 KB, 303 tokens by cl100k_base, as published. Nobody here has run it
Loop: Review, Patch, Retest
Iterate on review findings for a change until the review is clean, bounded by a hard iteration cap. Each iteration reviews, patches one finding cluster, and retests.
Loop Steps
- Review the change for correctness and contract impact; run
review-changefor the full pass. - Patch the highest-priority findings with a focused change.
- Rerun the relevant tests and checks and confirm the finding is resolved.
Max Iterations
The loop runs at most 3 iterations. When it reaches 3 iterations without meeting a stop condition, it stops unconditionally and reports the unfinished state and the outstanding work; it never raises the bound to keep going.
Stop Conditions
Stop the loop as soon as any of these holds:
- The relevant tests and checks are green.
- An iteration produces no diff.
- The same failure repeats identically across two consecutive iterations.
Approval Gate
- Get explicit human approval before any write, commit, or destructive step in each iteration.
- The loop never self-approves, never continues past the iteration bound, and never runs destructive commands on its own authority.
- Pause and surface the state whenever approval is missing.
Safety
- Do not upload source code.
- Do not read or print secrets.
- APC does not run this loop; a human or agent follows these instructions and remains in control.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.
Gives 0 of the 12 instructions most review quality skills give in 303 tokens
Counted across 1,048 of the 1,783 authors here whose files we hold, read 2026-08-07
- Ask questions one at a timein 81 of 1048, across 64 files
- Provide a recommended answer for each questionin 73 of 1048, across 50 files
- Explore the codebase instead of asking answerable questionsin 66 of 1048, across 42 files
- Resolve dependencies between decisions one-by-onein 42 of 1048, across 17 files
- Interview the user relentlessly about the planin 38 of 1048, across 13 files
- Order findings by severityin 31 of 1048
- Resolve each branch of the decision treein 27 of 1048, across 5 files
- Run a grilling sessionin 26 of 1048, across 5 files
- Update CONTEXT.md immediately when a term is resolvedin 26 of 1048, across 11 files
- Propose precise canonical terms for vague languagein 25 of 1048, across 7 files
- Create documentation files lazilyin 24 of 1048, across 5 files
- Assign severity to every findingin 24 of 1048
Said here and by no other author read
- patch the highest-priority findings
- rerun the relevant tests and checks
- confirm the finding is resolved
- report the unfinished state and outstanding work
- get explicit human approval before writes or destructive steps
- pause and surface state when approval is missing
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.