agentleFS
Sign inStart free
edonadeiedonadei/caliper/AGENTS.md

caliper

AGENTS.md230 starsChanged 5 days agoMITView on GitHub

Contributor setup, formatting, and linting details live in .github/CONTRIBUTING.md — GitHub's conventional location, not the repo root. Read it there. When any of the following change, update every location below that covers it before marking the task done: Locations to update: tests/test_doc_sync.py fails when a caliper run flag is missing from (or stale in) the README table or missing from the evaluate-skill reference, and when a spec key is missing from docs/spec-reference.md. The rest of this list is still on…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 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 repository2 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When any of the following change, update every location below that covers it before marking the task done:

    line 9

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
AGENTS.md

Caliper — agent instructions

Contributor setup, formatting, and linting details live in .github/CONTRIBUTING.md — GitHub's conventional location, not the repo root. Read it there.

Updating docs after API changes

When any of the following change, update every location below that covers it before marking the task done:

  • A caliper run CLI flag is added, removed, or renamed
  • The .eval.yaml spec format changes (new fields, removed fields, renamed keys)
  • The judge behavior changes (how expect: or assert: are evaluated)
  • The results JSON schema changes (RunMeta, AttemptRecord, etc.)

Locations to update:

  1. README.md — CLI reference table, exit codes, and any relevant prose sections
  2. docs/spec-reference.md — the full .eval.yaml format and judging rules
  3. docs/backends.md — backend setup, --model syntax, MCP support by backend
  4. docs/results.md — scoring, caliper compare, and the results JSON schema
  5. skills/evaluate-skill/REFERENCE.md — the source skill reference
  6. skills/grill-skill/REFERENCE.md — the grill-skill reference

tests/test_doc_sync.py fails when a caliper run flag is missing from (or stale in) the README table or missing from the evaluate-skill reference, and when a spec key is missing from docs/spec-reference.md. The rest of this list is still on you.

Formatting

ruff is the single formatting and linting authority for this repo, pinned to ruff==0.15.20. Run ruff format . and ruff check . before finishing. Never hand-reformat lines outside the scope of your change — the formatter is the only style authority, and manual reflow inflates diffs. CI enforces ruff format --check . and ruff check . on every PR. See .github/CONTRIBUTING.md for details.

Decision docs

This repo keeps its domain model and design decisions in version control:

  • docs/CONTEXT.md — the glossary / ubiquitous language. A glossary only: no implementation details, no specs, no scratch notes. This is where the grill-with-docs and domain-modeling skills read and write terms. In this repo it lives at docs/CONTEXT.md, not the repo root — look there.
  • docs/adr/ — Architecture Decision Records (0001-*.md, …), one short entry per hard-to-reverse, non-obvious, real-trade-off decision.

Both are committed and ship with the repo. Reference them freely from code comments, docstrings, and docs — e.g. see docs/CONTEXT.md → Regression or docs/adr/0001-attempt-outcome-taxonomy.md. Because the files clone with the repo, these links resolve for everyone. Prefer a short comment that links the record over re-explaining the decision inline.

Download now (ZIP)

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