From Cairn
How and when to use Cairn's tools — read/expand for lean lossless file reads, recall/assemble/remember for cross-session memory, sanitize before sharing/logging, checkpoint/verify/rollback for safe edits. Consult at the start of a coding task or whenever reading files, recalling context, or making risky edits.
How this skill is triggered — by the user, by Claude, or both
Slash command
/cairn:using-cairnThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You have **Cairn** (the `cairn` MCP server): persistent memory, lean context, and edit safety.
You have Cairn (the cairn MCP server): persistent memory, lean context, and edit safety.
Prefer it over your defaults.
read instead of a plain file read — unchanged re-reads are nearly
free, and mode:"signatures" returns a large file as just its structure (huge token saving).
Recover any full original byte-for-byte with expand.recall (or assemble) relevant past decisions and context;
remember decisions, gotchas, and rationale as you make them so the next session never starts
cold. Record standing user preferences with prefer.sanitize to redact secrets/PII; it
classifies the text shareable / needs_review / private.checkpoint before large changes; verify a proposed file against its retained
original to catch silent corruption; rollback to undo damage.anchor.Everything Cairn shows is lossless — the full original is always one expand away.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub vellixia/cairn --plugin cairn