The Agent Skills Directory

[SAFE]: The scripts/codebase_analyzer.py script is implemented using only the Python standard library. It performs read-only filesystem operations to identify directory structure and file types without executing external code or performing network requests.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes file and directory names from repositories, which serves as an ingestion point for untrusted data. If a scanned repository contains file names crafted with malicious instructions, these strings would be included in the analysis output provided to the agent.\n
Ingestion points: Repository filenames and directory names are read from the user-specified path by the analyzer script.\n
Boundary markers: The script output does not currently implement specific boundary markers or delimiters to isolate filenames.\n
Capability inventory: The skill can read file metadata but lacks network access, file write capabilities, or command execution functionality.\n
Sanitization: There is no filtering or sanitization of filenames within the script's output to neutralize potential instruction strings.

codebase-onboarding