Enables AI coding agents to consult persistent local project memory before acting, checking proposals against rejected ideas, repeated failure loops, goal drift, stale assumptions, and evidence-based completion criteria, and returning ALLOW, WARN, or BLOCK decisions so context loss after compaction doesn't cause repeated mistakes.
Stops your AI from re-introducing bugs, leaking provider keys, or weakening tests. Bug fixes become permanent regression guards; blocked mistakes become AI lessons the agent reads and learns from before its next edit.
Enables AI coding agents to avoid irreversible mistakes by intercepting risky deletions, overwrites, and commands before execution, applying path, pattern, and blast-radius safeguards with quarantine backups.
A runtime gate for coding agents. Blocks the tool calls that wreck a repo (force-push main, rm -rf, secret exfiltration, CI wipe) and lets normal build and commit work through. Machine-checked git-branch core (z3); the rest is high-precision heuristics. Tested on 3,790 real CI commands, 0 false blocks.
Acts as a production-grade safety layer for AI-assisted coding, monitoring Git hygiene, scanning for security issues (PII, secrets, injection), and enabling semantic history search.