precedent
Checks an AI coding agent's "I can't" and "done" claims against evidence.
- tools
- 5
- transport
- STDIO
- last updated
- Nov 13, 2026
Tools
5 toolsPublic tool metadata for what this MCP can expose to an agent.
find_precedentSearch past agent sessions, shell history and git for an action that previously worked on a similar task.
Search past agent sessions, shell history and git for an action that previously worked on a similar task.
No parameter schema in public metadata yet.
review_replyChecks an agent's reply for give-ups without evidence and unchecked 'done' claims.
Checks an agent's reply for give-ups without evidence and unchecked 'done' claims.
No parameter schema in public metadata yet.
record_failureRecords a failed attempt so repeated errors are detected and the agent is told to change route.
Records a failed attempt so repeated errors are detected and the agent is told to change route.
No parameter schema in public metadata yet.
check_test_editChecks a proposed edit for test weakening such as skip, xfail, .only, assert True, removed assertions or deleted test files.
Checks a proposed edit for test weakening such as skip, xfail, .only, assert True, removed assertions or deleted test files.
No parameter schema in public metadata yet.
check_commandChecks a shell command against the guard rules for destructive or corner-cutting operations.
Checks a shell command against the guard rules for destructive or corner-cutting operations.
No parameter schema in public metadata yet.
Sponsors
Agent CaffeineSponsoredClose your lid, come back to finished work
SnugSponsoredThe clipboard manager for the AI eraPromote your product hereReach 125k+ visitors monthly
WaxSponsoredBlock anything with supervision, without erasing your iPhonePromote your product hereReach 125k+ visitors monthly
knobsSponsoredSee your site on any phone, any day, in any languageReadme
precedent
Stop false "I can't"s and unchecked "done"s in your AI coding agents.
Enforced automatically in Claude Code, Codex CLI, Gemini CLI, OpenCode and GitHub Copilot CLI; available as MCP tools plus rules in Cursor, Windsurf, VS Code and Claude Desktop.
You've seen this: the agent says "I can't do that", you reply "you did it before", and suddenly it works. The capability was there all along. What it lacked was a precedent.
precedent finds that precedent automatically, from all your agents'
history, before the agent gives up. It also makes sure a pushed agent never
fakes success to please you.
Before / after
Without precedent:
You: convert report.xlsx to csv
Agent: Sorry, I can't convert Excel files in this environment.
With precedent, the agent's give-up is sent back with this (real hook output):
Details
- Repository
- github.com/juw12/precedent
- Website
- github.com/juw12/precedent
- Category
- AI & LLM Tools
- Package
https://github.com/juw12/precedent/releases/download/mcp-v3.3.0/precedent.mcpb- Transport
- STDIO
- Registry
- MCP registry, active
More from juw12
Related AI & LLM Tools MCP servers
View all
Duplicati Mcpkcofoni/duplicati-mcpMCP server for Duplicati backup management — list, run, monitor and configure jobs from an LLM.0stars
Kordi — Subscription Intelligencekordi-labs/kordi-mcp-serverTrack, analyze, and act on your streaming and SaaS subscriptions from any AI agent.0stars
maviapikuzenlabs/maviapi-mcpLive JSON from sites with no API (Booking, Airbnb, Tripadvisor…), plus agent-built APIs for any site0stars
Canarilanbass869-cell/canariSigned, offline-verifiable attestations for agent components; ALLOW/BLOCK/REVIEW per install.0stars
Invovate Mcp Serverlightspeedplusone/invovate-mcp-serverAI-agent invoice generator: PDF, JSON & UBL in 11 languages via the Invovate API.0stars
gnomonlihtness/gnomon-mcpDeterministic batch tools so LLM agents stop guessing dates and math.0stars