Releases: epfl-lara/LeanProbe
Release list
LeanProbe 0.4.0
- Split the agent contract out of
AGENTS.mdinto an installable skill at
src/lean_probe/skill/SKILL.md(YAML frontmatter + the MCP tool contract). - New
lean-probe install-skilldrops it into agent clients' skills
directories —~/.claude/skillsand~/.codex/skillsshare the same
skills/<name>/SKILL.mdlayout. Default installs to every present client
("both if both");--client claude|codexforces one,--skills-dir PATH
targets a root,--dry-run/--printpreview. The skill ships inside the
wheel, sopip install lean-probe && lean-probe install-skillneeds no
repo checkout. - A plain
pip install lean-probestays skill-free — the skill is purely
opt-in and a regular install writes nothing to~/.claude/~/.codex. - The README links to the skill as the single source of truth; a test keeps its
tool table in sync withTOOL_NAMES.
LeanProbe 0.3.1
LeanProbe 0.3.1
Fast Lean 4 proof feedback for AI agents — an MCP server, CLI, and Python API,
built on LeanInteract. It keeps a Lean REPL warm and reuses the elaborated
environment, so repeated checks return in tens of milliseconds instead of the
seconds a fresh lake build costs (~9–14× faster for sequential same-file work).
It never edits files; lake build stays the final whole-project gate.
Install (MCP server now included)
pip install lean-probe # or: uvx lean-probe mcp
claude mcp add lean-probe --env LEAN_PROBE_AUTO_BUILD=0 -- lean-probe mcpWorks with Claude Code, Codex, and any MCP client (config snippets in the README).
What it gives an agent:
- lean_check — verify any standalone snippet (the default; no file/name needed).
- lean_check_target — check or replace a declaration in a project file against its warm prior environment (sub-second after the first call).
- lean_proof_state / lean_tactic / lean_close_proof — drive a sorry tactic by tactic.
- Agent-first: ships usage instructions on connect, actionable error_code + hint
on every failure, and a compact inline feedback_lean for the next attempt.
Since 0.3.0
- mcp is now a core dependency — pip install lean-probe ships the server (no extra).
- Renamed AGENT.md → AGENTS.md (using-the-tool contract + contributor guide).
- Quickstart-first README with copy-paste setup for Claude Code, Codex, and generic clients.
0.3.0 was the agent-usability redesign: server instructions, the lean_* tool set
with the low-friction lean_check, a guarded replacement, env-aware REPL recovery,
and the compact feedback_lean. Full notes in CHANGELOG.md.
LeanProbe 0.2.2
What's Changed
- Added
lean_probe_capabilitiesMCP tool for reporting LeanProbe readiness, project-root detection, selected REPL path, and live session state. - Added
lean-probe capabilitiesCLI command with matching readiness output. - Documented how LeanProbe complements LSP-backed Lean MCP tools such as
lean-lsp-mcp. - Updated tests for the new MCP and CLI capability surfaces.
Validation
python -m pytest -qpython -m ruff check .python -m buildpython -m twine check dist/*
LeanProbe 0.2.1
Patch release for LeanProbe MCP setup docs.
- Documents that stdio MCP clients should keep
LEAN_PROBE_AUTO_BUILD=0. - Explains that Lean/Lake projects should be built before starting LeanProbe MCP.
- Prevents MCP transport corruption from build output written to stdout.
Install:
python -m pip install -U lean-probeMCP:
python -m pip install -U "lean-probe[mcp]"
lean-probe mcpLeanProbe 0.2.0
First public LeanProbe release.
Install:
python -m pip install lean-probeMCP support:
python -m pip install "lean-probe[mcp]"
lean-probe mcpHighlights:
- Standalone Python package, CLI, and MCP server for LeanInteract-backed Lean feedback.
- Cached same-file declaration checks.
- Diagnostics, tactic metadata, proof-state stepping, and feedback_lean.
- Benchmarks for repeated target checks and sequential same-file checks.
- Python 3.10-3.13 support across Linux, macOS, and Windows.