Invalidate an agent memory when its supporting file changes
At a glance
Anish builds Mnemolith to stop coding assistants from relying on facts made stale by code changes. This workflow is listed under agent-memory, cache-invalidation, git, developer-tools.
The MCP server shares project memory across coding tools and links memories to supporting files' Git blob hashes. Changed contents set memories aside for review; a dashboard lets a developer confirm or retire them.
How it went
In the author's calibration, a Postgres-to-MySQL contradiction scored only 0.30 similarity, motivating agent judgment for contradictions.
Worth knowing
The reported version is 0.2.1; team sharing is limited to synchronized AGENTS.md or CLAUDE.md sections.
Save ideas in this browser. Upvotes count across the board.
Open-source local procedural memory that learns only after tests pass, then hands Claude Code/Codex/Cursor a short verified skill note on similar tasks (reported token/tool cuts in published benchmarks).
Show HN (Oct 17): RugSnare pins each MCP tool's {name, description, inputSchema} by canonicalized SHA-256 at install, then diffs upgrades across 24 popular servers (~3.6M weekly installs)—336 contract changes found; grades BREAKING/LOOSENED/EXPANDED/NOTATION/COSMETIC and exits 1 for CI. Apache-2.0, zero npm deps; OWASP MCP Security Cheat Sheet now recommends contract pinning.
Local-first OSS DB/workspace where Claude Code/Codex/Cursor share durable records, docs, skills, and apps—writes that matter go through human review; Skill, MCP, OpenAPI, and ACP connections keep agents stateless and work durable.