The Agent Skills Directory

[SAFE]: The skill performs local file indexing and searching of documentation as advertised. The Python scripts strictly use standard libraries for text processing and JSON handling. No evidence of data exfiltration, remote code execution, or persistence mechanisms was found. All analyzed components align with the stated purpose of providing a developer tool for Chromium documentation.- [INDIRECT_PROMPT_INJECTION]: The skill processes content from external documentation files which could contain embedded instructions.
Ingestion points: Document indexing scans markdown files in the source tree using the chromium_docs.py script.
Boundary markers: Search results are presented with structured markdown headers and links, providing structural separation between results.
Capability inventory: File system access is limited to reading documentation files and writing index files to the skill's data directory.
Sanitization: Standard markdown parsing is used, and content is summarized for display without additional sanitization filters.

chromium-docs