The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several git commands (log, diff, status, rev-parse, cat-file) to analyze repository changes and determine which vault notes need updating based on source code modifications.\n- [DYNAMIC_EXECUTION]: Executes a local Node.js script (vault-stale.mjs) included in the skill package. This script reads vault note frontmatter and uses git diff to identify files that have changed since their last documented synchronization.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically git commit messages and the current conversation history, to summarize decisions and findings for the vault.\n
Ingestion points: Commits retrieved via git log and the current interaction history (SKILL.md).\n
Boundary markers: Instructions mandate following existing vault conventions and require the agent to cite repository paths and commits for all claims.\n
Capability inventory: The agent has permissions to write to specific vault markdown files and execute local git commands and the vault-stale.mjs script.\n
Sanitization: The skill includes explicit instructions for the agent to avoid recording secrets, keys, or personal data, and limits file writing to a pre-configured vault directory.

vault-sync