agentleFS
Sign inStart free
r-irber-irbe/proof-skills/skills/lean-setup/SKILL.md

lean-setup

Skill3 starsChanged 24 days agoApache-2.0View on GitHub

USE FOR: bootstrapping a Lean 4 project or toolchain, configuring elan, pinning `lean-toolchain`. DO NOT USE FOR: building Lake projects (@lean-build); bisecting regressions (@lean-bisect). TRIGGERS: elan, lean-toolchain, lake serve, moreServerArgs.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Create a free account to mount
Files
SKILL.md

lean-setup

MANDATORY (hot-tier): the gates in Section Behavioural rules and the Persist

Routing

  • USE FOR: configuring elan toolchains; pinning lean-toolchain; configuring Lake server arguments (moreServerArgs) to prevent OOM SIGKILL and stack exhaustion; discovering Lean C sysroot paths via lean --print-prefix; verifying lean --version and lake env lean --version agree; bootstrapping or repairing local builds.
  • DO NOT USE FOR: building Mathlib or downstream Lake projects (use @lean-build); bisecting which commit changed behaviour (use @lean-bisect); minimising an error into a reproducer (use @lean-mwe); writing proofs against an existing toolchain (use @lean-proof); scaffolding new lakefiles for

Workflow

  1. Discover [discover] -- inspect clone state: check lean-toolchain, elan list, and lakefile configuration. Identify whether this is fresh project setup, version bump, or toolchain repair.
  2. Plan [discover] -- choose toolchain release or disambiguated name; identify required server arguments (moreServerArgs). STOP if confidence < 80% on naming or version target.
  3. Execute [execute] -- install/link toolchain (elan toolchain install or cmake + elan toolchain link); write lean-toolchain; configure moreServerArgs in lakefile.lean.
  4. Verify [validate] -- lean --version matches target (G-6); lake env lean --version agrees (G-7); lake serve starts without watchdog errors. On disagreement, fix overrides; max 3 attempts then escalate.

Recovery & STOP

  • make -j -C build/release or elan download fails x3 -> STOP, escalate to human; do not retry blindly.
  • lean --version and lake env lean --version disagree after 3 fix attempts -> STOP, escalate (G-8 forbids handing off).
  • Scope drift (edit files outside lean-toolchain family or lakefile) -> immediate STOP, re-anchor.
  • Confidence < 80% on toolchain naming or release compatibility -> STOP, ask.

Handoffs

  • Predecessors / successors: see FM handoffs. lean-setup is a root node. Typical outbound: @lean-build (compile project), @lean-proof (start tactic proving), @lean-bisect (investigate regressions), @lean-mwe (reproduce toolchain bugs).
  • Sister skills: @lean-build for Lake builds; @lean-proof for live proving.
  • Source spec: specs/lean/setup/requirements.md.

Detailed reference

Full encyclopedic reference, decision trees, and lookup tables live in references/lean-setup-handbook.md .

See also

  • ../../references/lean-setup-handbook.md -- Complete lean-setup Handbook
  • ../lean-gateway/SKILL.md -- Top-level skill router

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