# Lemma MCP

> Drive the Lemma MCP from the terminal. Use when the user asks about Lemma issues, agent failures, incidents, or regressions; wants to review, validate, confirm, dismiss, fix, or resolve detections; or asks questions like "are there issues where users get frustrated", "any urgent lemma issues", "look through all urgent issues and validate each", or "dismiss everything under X impact." Also use when browsing, ranking, or acting on Lemma detections over MCP. Do not use for recording what an agent does in Lemma, or for writing agent context and artifacts from a codebase — that is lemma-artifacts.

- Skill: `uselemma/lemma-mcp` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add uselemma/lemma-mcp`
- Raw SKILL.md: https://api.skillmd.com/api/skills/uselemma/lemma-mcp/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: uselemma (https://skillmd.com/u/uselemma)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/uselemma/lemma-mcp

---


# Lemma MCP

You are driving the Lemma MCP. Lemma detects semantic failures in production
AI agents. Human verdicts on those detections become ground-truth training
signal that sharpens detection. That is why the rules below are strict about
who decides.

## Core rules

1. Reads run freely. Never call a write tool (`set_issue_close_feedback`,
   `set_issue_status`, `set_issue_feedback`) without the user's explicit
   choice in this conversation. Propose verdicts; never write them on your
   own judgment.
2. Skip records nothing. Confirm, dismiss, and resolve always write back.
3. Never show the user a raw UUID or raw JSON. Render issues as an 8-char id
   prefix plus a list ordinal, e.g. `[2] a3f81c92`. Resolve references like
   "the second one" yourself.
4. Copy style in everything you render: sentence case, terse, no em-dashes.

## Workflows

| Area | When | Reference |
| --- | --- | --- |
| Issues | Triage detections: list, rank, drill into evidence, confirm / dismiss / resolve | [references/issues.md](references/issues.md) |

Read the matching reference before acting. Keep tool names exactly as written
there (`list_issues`, `get_issue_chat_context`, `set_issue_close_feedback`,
and related MCP tools) so transport mappings stay valid.