Cairn Context Continuity-ecb3bf — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Cairn Context Continuity-ecb3bf (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.
Cairn is a local MCP server that keeps decisions + evidence across sessions/models and warns before you run into the degraded part of the context window. Use its four tools:
surface caps the window) surfaceCap. It returns the zone (green/yellow/red) relative to the active window. Check it when the session feels long. The Claude Code statusline shows this ambiently; call the tool when you want the exact boundaries.
context_status is yellow or red, or right beforea compaction, produce a 7-bucket brief and call handoff with it (account mode — you, the host agent, write the brief on the user's own subscription; Cairn just freezes verbatim spans and persists). The 7 buckets: DECISIONS (mark superseded ones), EVIDENCE (with source refs), OPEN QUESTIONS, CONSTRAINTS, VERBATIM (code/IDs/values byte-exact), NEXT STEPS, DISCARDED. Pass requiredVerbatim for any value that MUST survive byte-exact, and structured decisions/evidence when you can extract them. Prefer producing the brief in a subagent (Task tool) so the compaction reasoning tokens don't bloat the main context (ADR-0007).
in place — after a host auto-compaction the SessionStart(compact) hook re-injects the brief faithfully (not summarized). For a fresh start or the other host, the ledger re-injects via decision_log / SessionStart.
view:"current" to reconstruct accepted + open decisions (superseded ones are kept but excluded). On resume this re-injected state is the source of truth — don't re-read the whole repo; open only the files named in NEXT STEPS. Use view:"all" for the full history + evidence.
cairn-resume) — re-inject the ledger and continue without re-reading the repo.cairn-handoff) — author + persist a 7-bucket brief now.cairn-help) — show the command / trigger reference table.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.