The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local files which could contain malicious instructions designed to influence the agent or the target knowledge systems.
Ingestion points: The skill reads content from CHANGELOG.md, CLAUDE.md, and src/hooks/README.md (Step 1).
Boundary markers: There are no explicit delimiters or instructions provided to the agent or the target tools to ignore potential instructions embedded within the release text.
Capability inventory: The skill has the ability to write to external services via mcp__notebooklm-mcp__source_add, mcp__hq-content__knowledge_ingest, and mcp__notebooklm-mcp__studio_create.
Sanitization: There is no evidence of sanitization, filtering, or validation of the file content before it is interpolated into tool calls and sent to external platforms.

release-sync