agentleFS
oborchersoborchers/proqi/xtask/AGENTS.md

proqi / xtask

AGENTS.md45 starsChanged 51 days agoMITView on GitHub

The repository root contract applies here. xtask is the sole project-owned command surface for development, CI, packaging, and release verification. - Keep subcommands thin, deterministic orchestrators around standard tools and repository policy. Do not add a shell script, Make target, or workflow-only implementation when xtask can own the operation portably. - quality is deliberately test-free. test owns the deterministic suite and check-full is exactly the canonical final aggregate of quality followed by test. check is the transparent iterative gate: ordinary…

The file

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

4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

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

    Keep subcommands thin, deterministic orchestrators around standard tools and repository policy.

    line 8
  2. Very specific:

    Every AGENTS.md at any depth has a sibling relative CLAUDE.md -> AGENTS.md symlink.

    line 20
  3. Very specific:

    Keep build, package, coverage, and review artifacts below ignored target/ or a securely created temporary directory.

    line 23
  4. Very specific:

    Treat GitHub Actions as scheduling and permissions configuration.

    line 33

≈ “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

Proqi engineering-operations contract

The repository root contract applies here. xtask is the sole project-owned command surface for development, CI, packaging, and release verification.

Command ownership

  • Keep subcommands thin, deterministic orchestrators around standard tools and repository policy. Do not add a shell script, Make target, or workflow-only implementation when xtask can own the operation portably.
  • quality is deliberately test-free. test owns the deterministic suite and check-full is exactly the canonical final aggregate of quality followed by test. check is the transparent iterative gate: ordinary code runs quality, doctests, and every non-PTY nextest binary; high-risk or ambiguous changes fail closed to check-full. CI jobs may split final phases but must not redefine or redundantly rerun the complete suite.
  • A policy check reports actionable paths and reasons, has accepted and rejected fixture coverage, and scans the repository independently of developer machine state. Never weaken a policy to accommodate one implementation.
  • Every AGENTS.md at any depth has a sibling relative CLAUDE.md -> AGENTS.md symlink. Instruction validation remains part of the canonical quality path.
  • Keep build, package, coverage, and review artifacts below ignored target/ or a securely created temporary directory. A verification command never publishes, installs globally, edits Git configuration, or changes repository settings without a separately authorized command.

Compatibility and verification

  • Preserve the pinned toolchain, MSRV, platform package contracts, source-only crate allowlist, Debian install/remove/reinstall contract, and exact artifact identity across release stages.
  • Treat GitHub Actions as scheduling and permissions configuration. Commands, policy, and package behavior stay locally executable through cargo xtask.
  • Tests for operations tooling must not require credentials, user configuration, the network, Docker, or a particular host path unless that capability is the explicit subject of a separately gated command.

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.