Install
$ agentstack add skill-merleanprover-merlean-init-merlean ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ● Environment & secrets Used
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
init_merlean — one-command MerLEAN setup
.claude/ (skills + agents + settings) is tracked in git, so there is nothing to "deploy": this skill prepares the runtime only. Everything is done by one idempotent, non-interactive script:
bash .claude/skills/init_merlean/setup.sh [--openai-key sk-...] [--skip-venv] [--skip-mathlib]
Procedure
- Key first. Check whether an OpenAI key is already reachable:
.envat the repo root
contains OPENAI_API_KEY=sk-..., or OPENAI_API_KEY is set in the environment. If neither, ask the user for their key (it is required by the Mem0-g plan graph for embeddings/rerank); they may also skip and fill .env in later.
- Run the script (from the repo root):
- key in hand →
bash .claude/skills/init_merlean/setup.sh --openai-key - key already reachable, or user skipped →
bash .claude/skills/init_merlean/setup.sh - The Mathlib step (
lake update+lake exe cache get+lake build) can take many
minutes on first run — run it with a generous timeout or in the background. Pass --skip-mathlib only if the user asks for a quick/venv-only setup.
- Read the check block the script prints at the end. All three must be
[ok]:
lean-lsp-mcp importable from the venv, lake on PATH, plan_store importable.
lake not found→ tell the user to install Lean viaelan
(https://leanprover-community.github.io/get_started.html), then re-run this skill.
- venv failures → the script already tried
ensurepip/get-pipand auvfallback;
report the pip error to the user.
- Restart note. If
.mcp.jsonwas just (re)generated, tell the user thelean-lspMCP
server loads on the next Claude Code session start.
What the script does (all idempotent)
| step | action | |---|---| | venv | build .venv/ from src/requirements.txt (host python ≥ 3.9, else uv fallback) | | .env | write from --openai-key, else preserve an existing key, else leave the template | | .mcp.json | generated from templates/mcp.json — registers the lean-lsp MCP server, run from the venv python | | Lean workspace | copy lean-toolchain / lakefile.toml / MerLeanExperiment.lean to the root if missing, then lake update + lake exe cache get + lake build (toolchain auto-synced to Mathlib's) | | checks | lean-lsp-mcp import, lake on PATH, plan_store import |
Existing root files (.env, lean-toolchain, lakefile.toml, MerLeanExperiment.lean) are never overwritten (only .mcp.json is regenerated). Re-running after a failed step is safe.
After setup, the plan-graph CLI is invoked as .venv/bin/python src/cli.py --data (Windows/Git Bash: .venv/Scripts/python.exe) — see the plan-graph skill.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: MerLeanProver
- Source: MerLeanProver/MerLean
- License: Apache-2.0
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.