Install
$ agentstack add skill-andreratzenberger-loopify-loopify-run ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
About
loopify-run
Execute a loop from an existing loop directory or Loop Contract.
When To Use
- The repo already has a
.loopify/loops/NNN-slug/directory, Loop Contract,
or quality gate.
- The user asks to run until checks pass.
- A goal should proceed by evidence rather than prose confidence.
Do Not Use
- Do not run outside the contract's authority boundaries.
- Do not keep patching after the blocked rule is met.
- Do not claim success without running the required checks.
- Do not write
verdict.md— the maker never grades its own homework.
Workflow
- Locate the loop directory, or accept a named Loop Contract path.
- Read
loop-contract.md. - Confirm allowed paths, denied paths, budgets, and stop rules.
- Run the loop directory
quality-gate.shor specified checks. - Inspect the exact failure.
- Patch the smallest useful cause.
- Update
trace.md. - Rerun checks.
- Repeat according to
references/loop-execution-protocol.md. - When checks pass and manual items are queued, request verification per the
contract's ## Verification section: hand contract, trace, diff, and gate output to an independent verifier (different model, fresh sub-agent, or human — see loopify-review verifier mode). Never write verdict.md yourself.
- On reject: append the findings to
trace.mdas observed failures and
continue within budget. If the verifier rejects the same finding twice, stop escalated. On cannot-verify: stop escalated.
- Write
final-report.mdwithMaker:identity and aVerdict:link. - Stop with
success(requires an approving verdict),blocked,
escalated, or budget-exhausted, and validate the claim with check-stop-reason.mjs .
Output
Updated artifacts, updated loop-directory trace, and a final report using templates/run-final-report.md.
Preferred loop path: .loopify/loops/NNN-slug/. Legacy direct contract paths are supported when the user names one.
Validation
- The final report names the stop reason.
- The trace records checks, failures, patches, and rationale.
- Passing checks are scoped to what they prove.
.loopify/index.mdis updated when the loop status changes.check-stop-reason.mjsexits 0 for the claimed stop reason.successclaims carry an approvingverdict.mdwhose verifier differs from
the final report's Maker:.
Example
Use loopify-run on .loopify/loops/001-papertrail/. Stop only when quality-gate.sh passes or the blocked rule is met.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: AndreRatzenberger
- Source: AndreRatzenberger/loopify
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.