The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local development commands including npm run preindex, git synchronization and commit operations, and gh (GitHub CLI) for pull request management. It also invokes a TypeScript build script using npx tsx scripts/build-catalog.ts to verify data consistency before staging changes.\n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection (Category 8) due to its core function of ingesting data from external repositories. Instructions or malicious content embedded in the refreshed repositories could influence the agent during the generation of the refresh summary or the Pull Request body.\n
Ingestion points: data/skill-index-resources.json, data/skill-index/*.json, and preindex.log (which captures output from processing external repositories).\n
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions when interpolating external repository names, counts, or error messages into the user-facing summary.\n
Capability inventory: The skill can perform shell command execution, file modifications, and network operations (via git push and gh pr create).\n
Sanitization: Absent. There is no evidence of validation or filtering for data fetched from the external sources before it is processed by the LLM.

refresh-index