agentleFS
superlinkedsuperlinked/sie/AGENTS.md

sie

AGENTS.md3.5k starsChanged 6 days agoApache-2.0View on GitHub

This repository is the source of truth for the open-source SIE server, SDKs, gateway, sidecar, configuration service, MCP service, Helm chart, and public release automation. Keep changes self-contained: a clean clone must build and test without access to another repository, private registry, or secret. Every pull request runs the public lint, typecheck, unit, integration, and package/container checks; these are not selected by changed paths. Benchmark and quality-evaluation jobs are not part of this CI. The separate Real models workflow serves…

Trending

More talk lately.about the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 11 mentionsMay 2026: 4 mentionsJun 2026: 1 mentionJul 2026: 0 mentionsAug 2026: 2 mentionsSep 2026: 4 mentionsOct 2026: 5 mentionsNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the superlinked/sie repository that our weekly crawl found. There is no adoption data for this file yet.

Who's talking about superlinked/sie

The biggest 5 of 27 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 27

Talked about on GitHub, Hacker News and Bluesky (27 times).

On GitHub · 21

16 more on GitHub

On Hacker News · 3

  • 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 31,439 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.

The file

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

10 rules. About 1 is common, 3 are distinctive, and 6 are very specific to this project.

  • ≈1

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Use ordinary shell tools for read-only inspection.

    line 12
  2. Very specific:

    Run project tasks with mise run <task> and pass task arguments after --.

    line 13
  3. Very specific:

    Run version-managed executables that are not tasks with mise exec --.

    line 14
  4. Very specific:

    Run mise tasks before assuming a task exists.

    line 16
  5. Distinctive:

    Every pull request runs the public lint, typecheck, unit, integration, and package/container checks.

    line 18
  6. Distinctive:

    Keep the root Python and pnpm locks authoritative.

    line 34
  7. Very specific:

    Use sie_sdk.SIEClient for SIE API examples.

    line 36
  8. Very specific:

    Keep __init__.py files empty and imports at module scope except for optional dependencies.

    line 38
  9. Common:

    Do not commit secrets, credentials, generated Helm dependencies, staged chart model/bundle files, build output, or package archives.

    line 40
  10. Very specific:

    Read RELEASE.md before authoring, publishing, or recovering a release.

    line 48

≈ “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

AGENTS.md — Public SIE contributor guide

Scope

This repository is the source of truth for the open-source SIE server, SDKs, gateway, sidecar, configuration service, MCP service, Helm chart, and public release automation. Keep changes self-contained: a clean clone must build and test without access to another repository, private registry, or secret.

Commands

  • Use ordinary shell tools for read-only inspection.
  • Run project tasks with mise run <task> and pass task arguments after --.
  • Run version-managed executables that are not tasks with mise exec --.
  • Bootstrap a clean checkout with ./tools/init.sh.
  • Run mise tasks before assuming a task exists.

Every pull request runs the public lint, typecheck, unit, integration, and package/container checks; these are not selected by changed paths. Benchmark and quality-evaluation jobs are not part of this CI. The separate Real models workflow serves pinned public checkpoints on CPU nightly, on manual dispatch, and for pull requests that change model-serving, SDK, or integration paths. It is advisory and not part of CI / Required.

The main tasks are mise run lint, mise run typecheck, mise run test, mise run test-integrations, mise run ts -- build, mise run ts -- lint, mise run rust-check, mise run rust-test, and mise run helm -- lint. Build the TypeScript workspace before checking dependent packages in isolation. The standalone Candle worker uses mise exec -- cargo <command> --manifest-path packages/sie_server_rust/Cargo.toml --locked.

Development boundaries

  • Keep the root Python and pnpm locks authoritative; standalone Rust crates keep their checked-in locks.
  • Use sie_sdk.SIEClient for SIE API examples. Use current gateway terminology; legacy wire names remain only where compatibility requires it.
  • Keep __init__.py files empty and imports at module scope except for optional dependencies.
  • Do not commit secrets, credentials, generated Helm dependencies, staged chart model/bundle files, build output, or package archives.
  • Trust-boundary changes to ingress, authentication, identity handling, billing, or wire protocols require an adversarial security review.
  • Use Conventional Commits.

Releases

Read RELEASE.md before authoring, publishing, or recovering a release. Terraform modules have independent release lifecycles: a new SIE version alone does not require module documentation updates or module releases. Follow Terraform module follow-up only when a release changes an infrastructure requirement or exposes a real compatibility issue. Documentation-only changes use non-releasing commits; reserve module versions for actual module changes. An explicitly authorized one-time Registry documentation refresh follows the exception in that guide; never retag an existing module version to change its documentation.

Release automation is intentionally fail-closed. Package, image, and chart builds may run without publication authority; external publication additionally requires the protected workflow inputs and repository publishing latch described in RELEASE.md. Never bypass version, tag, source-revision, or full-set verification.

Release-PR authoring is independently default-off and requires the repository variable PUBLIC_RELEASE_AUTOMATION_ENABLED to be exactly true. Configure the release App and protected release-automation environment before enabling it; do not confuse this authoring gate with artifact publication enablement.

The release-please baseline is 0.7.3 at exact public commit 60996d9c30168e0f8e85b680295f147fdee87f61. Keep that full SHA as the checked-in bootstrap boundary; do not create a substitute baseline tag or move the boundary to a later commit. Publication retries reuse the original release run and artifacts, not a newer-main rebuild.

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.