agentleFS
Sign inStart free
CultureDigitaliCultureDigitali/gigiloop/.cursor/rules/gigiloop.mdc

gigiloop / rules

Cursor rule0 starsChanged 50 days agoMITView on GitHub

Verification-first autonomous coding loop with resumable state, adversarial review, and local CI fallback

We split this gigiloop.mdc into the rules it gives an agent

9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use GigiLoop when the user asks for repeated autonomous improvement rather than a one-pass response.

    line 7
  2. Distinctive:

    keep iterating until it works.

    line 10
  3. Distinctive:

    do not stop until done;

    line 11
  4. Distinctive:

    keep testing and refining.

    line 14
  5. Very specific:

    use gigiloop/scripts/gigiloop.py when Python/shell are available.

    line 21
  6. Distinctive:

    use Builder, Verifier, Red Team, Judge, and optional Improver separation when subagents/fresh contexts are available.

    line 24
  7. Distinctive:

    run targeted checks during iterations and broad checks at milestones/final gate.

    line 25
  8. Distinctive:

    if GitHub Actions/hosted CI is quota-exhausted or unavailable, continue with equivalent local checks while recording the missing remote …

    line 31
  9. Very specific:

    Read gigiloop/references/runtime.md, gigiloop/references/orchestration.md, and gigiloop/references/integrity.md when their rules apply.

    line 36

≈ “Common” is an estimate: at least 1 in 5 of the 6,350 Cursor rule files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
gigiloop.mdc
---
description: Verification-first autonomous coding loop with resumable state, adversarial review, and local CI fallback
globs: ["**/*"]
alwaysApply: false
---

Use GigiLoop when the user asks for repeated autonomous improvement rather than a one-pass response.

Typical triggers:
- keep iterating until it works;
- do not stop until done;
- Ralph-style loop;
- harden this for production;
- keep testing and refining;
- resume after interruption;
- improve the project after correctness is stable.

Protocol:
1. choose strict, balanced, or fast; default to balanced and never silently downgrade;
2. baseline the repository, pre-existing failures, verification contract, and uncommitted user work;
3. use `gigiloop/scripts/gigiloop.py` when Python/shell are available; canonical runtime state is `.gigiloop/checkpoint.json`;
4. define measurable acceptance criteria;
5. implement the highest-impact confirmed improvement;
6. use Builder, Verifier, Red Team, Judge, and optional Improver separation when subagents/fresh contexts are available;
7. run targeted checks during iterations and broad checks at milestones/final gate;
8. score only with current evidence tied to the current repository fingerprint;
9. red-team the diff, tests, assumptions, thresholds, and scope;
10. reconcile scores after critique;
11. reject deleted/skipped tests, lowered thresholds, disabled checks, blind snapshot approval, or narrowed verification used merely to obtain green output;
12. preserve unrelated user edits and avoid destructive operations without authorization;
13. if GitHub Actions/hosted CI is quota-exhausted or unavailable, continue with equivalent local checks while recording the missing remote evidence;
14. on resume, reconcile checkpoint and repository fingerprint before reusing evidence;
15. deliberately change strategy after a plateau;
16. exit as SUCCESS, BLOCKED, BUDGET EXHAUSTED, or STOPPED with an evidence-backed report.

Read `gigiloop/references/runtime.md`, `gigiloop/references/orchestration.md`, and `gigiloop/references/integrity.md` when their rules apply.

Download now (ZIP)

  • Where files are headingNew Cursor rule files have more rules about planning and tracking tasks, and fewer about code style.

    Of the Cursor rule files we read, the 3,978 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repositoryOne other file it gives its agents.

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free