agentleFS
Sign inStart free
glinckerglincker/stacklit/AGENTS.md

stacklit

AGENTS.md131 starsChanged 7 days agoMITView on GitHub

Guide for agents and people working on the Stacklit repo itself. If you only want to use Stacklit on your own project, read README.md and docs/INSTALL.md. A Go CLI that indexes a codebase into stacklit.json (modules, exports, dependency edges, hints) plus a symbol graph, and serves it over MCP. Deterministic, no LLM, no network calls from the CLI. Module path github.com/glincker/stacklit, Go version from go.mod. CI runs vet and tests on Linux, macOS and Windows. Node is only needed for…

  • Reads credentials
  • 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.

  • In this repository2 other files it gives its agents.

The file

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

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    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. Very specific:

    Commit subjects: <type>: <description> with type one of feat, fix, perf, refactor, docs, test, chore.

    line 38
  2. Distinctive:

    No em dashes or en dashes in code, comments, docs or commit messages.

    line 39
  3. Very specific:

    Add a failing test first for behavior changes.

    line 40
  4. Distinctive:

    Do not describe unmerged features as shipped.

    line 42
  5. Very specific:

    Do not commit generated files (stacklit.json, docs/DEPENDENCIES.md) unless the change is about them.

    line 43
  6. Very specific:

    Never commit secrets or .env files.

    line 45
  7. Very specific:

    Read stacklit.json before broad searches.

    line 54

≈ “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

Guide for agents and people working on the Stacklit repo itself. If you only want to use Stacklit on your own project, read README.md and docs/INSTALL.md.

What this is

A Go CLI that indexes a codebase into stacklit.json (modules, exports, dependency edges, hints) plus a symbol graph, and serves it over MCP. Deterministic, no LLM, no network calls from the CLI. Module path github.com/glincker/stacklit, Go version from go.mod.

Build and test

make build                 # ./stacklit
go vet ./...
go test ./... -race -count=1
go build -o /tmp/stacklit ./cmd/stacklit   # try it on another repo

CI runs vet and tests on Linux, macOS and Windows. Node is only needed for npm/.

Where things live

PathHolds
cmd/stacklitmain package
internal/cliCobra commands
internal/parser, internal/graphextractors and the dependency and symbol graph
internal/engine, internal/renderer, internal/schemapipeline, outputs, index types
internal/mcpMCP server (stdio)
internal/setupstacklit setup targets, hooks, the embedded skill
skills/stacklit/SKILL.mdthe skill; internal/setup/skill/SKILL.md and plugins/stacklit/skills/stacklit/SKILL.md must match it
plugins/stacklit, .claude-plugin/Claude Code plugin and marketplace
server.json, npm/MCP registry entry, npm wrapper
scripts/install.sh, the llms-full.txt generator, sync helpers
docs/ARCHITECTURE.mdpipeline detail

Rules

  • Commit subjects: <type>: <description> with type one of feat, fix, perf, refactor, docs, test, chore. Keep them short.
  • No em dashes or en dashes in code, comments, docs or commit messages. Plain sentences, no marketing language.
  • Add a failing test first for behavior changes. Parser tests sit next to the parsers and use testdata/.
  • Claims about size or speed in docs must come from bench/ output.
  • Do not describe unmerged features as shipped.
  • Do not commit generated files (stacklit.json, docs/DEPENDENCIES.md) unless the change is about them; CI refreshes them.
  • Release files (.goreleaser.yml, .github/workflows/release.yml) and npm/package.json are owner-controlled. Ask before editing.
  • Never commit secrets or .env files.

Keeping copies in sync

  • Skill copies: scripts/sync-plugin-skill.sh --check (run without --check to fix).
  • llms.txt is hand written. site/public/llms-full.txt is generated: sh scripts/gen-llms-full.sh, and --check fails if it is stale.

Using Stacklit on this repo

Read stacklit.json before broad searches. stacklit query <name> finds a symbol, stacklit impact <name> shows what depends on it.

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