kage-memory — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited kage-memory (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Kage gives this repo a memory that can be trusted: it rejects hallucinated citations on write, withholds stale memory on recall, and grounds knowledge to the code graph. Memory lives as plain files in .agent_memory/, reviewed in the same PR as the code.
npm install -g @kage-core/kage-graph-mcp
kage init --project .
kage setup <your-agent> --project . --write # claude-code, codex, cursor, windsurf, …Try it instantly, no setup: npx -y @kage-core/kage-graph-mcp demo
repo-specific decision, call kage_context with the repo path and the task as the query. Use the returned memory only when it's relevant and source-backed. Prefer repo memory over your assumptions.
something was withheld, don't go hunting for it; it's withheld because its evidence is gone.
convention or gotcha, or learn how to run/test/build the repo, call kage_learn with a concise learning and the file paths it concerns. Citations to files that don't exist are rejected — cite real paths.
describes, update or supersede that memory (kage_verify_citations, kage_compact) before finishing.
Prove it: kage benchmark --trust --project .
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.