encero-systems/incan / .agents/skills/loop
loop
VerifiedOrchestrate a named multi-skill loop such as /loop /review /fix until clean or blocked. Use when the user explicitly asks to loop skills, rerun a review/fix cycle, or says /loop with one or more skill names.
Loop — Skill Orchestrator
Purpose
/loop is a thin orchestrator. It does not define domain checks itself. It repeatedly runs a detector skill and a repair skill until the detector reports no remaining blockers/warnings or the work becomes blocked.
The loop should preserve state in the current worktree's persistent report file:
.agents/state/review-report.md
Typical use:
/loop /review /fix
That means:
- run
/review, - if it reports actionable findings, run
/fix, - run
/reviewagain, - repeat until clean or blocked.
Contract
- The first named skill should be a detector/reviewer.
- The second named skill should be a repair/fix skill.
- If the user provides more skills, run them in the given order, then loop back to the first detector.
Stopping conditions
Stop the loop when one of these is true:
- the detector reports no blockers or warnings,
- remaining findings are explicitly classified as:
- out of scope,
- risky without user confirmation,
- external blocker,
- separate compiler bug,
- the user interrupts or redirects the work.
Loop discipline
- Keep each cycle concrete: detector output, repair pass, detector rerun.
- Do not let a detector silently become a fixer or a fixer silently become a detector. Respect each skill's role.
- Use
.agents/state/review-report.mdas the durable state between cycles. Do not rely on conversational memory to remember which findings are still open. - Require the report to stay structured on every cycle:
## Scope,## Activity,## Files, and## Verificationmust remain present. A rewritten prose summary is not valid loop state. - Require file states to be honest loop state: files start
pending, then move toclean,findings,fixed, orblockedonly as the checklist evidence justifies. - When
/loopstarts a new detector-first run, the detector should initialize a fresh report scaffold for that run. Preserve state across cycles within the same loop, not across unrelated invocations. - If a broad verification command fails for unrelated reasons, continue the loop when local findings are still fixable.
- Summarize each cycle briefly so the user can see whether the loop is converging.
Output format
Produce a compact progress report:
## Loop — <skills>
### Cycle 1
<detector result>
<fix result>
### Cycle 2
...
### Final state
<clean / blocked / needs user direction>
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 465k | Updated | Claude CodeCodex | Unverified | Install |
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 465k | Updated | Claude CodeCodex | Unverified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 318k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 318k | Updated | Claude CodeCodex | Verified | Install |