The Agent Skills Directory

[SAFE]: No malicious patterns, data exfiltration, or dangerous command execution were detected. The skill facilitates legitimate codebase exploration through provided MCP tools.
[PROMPT_INJECTION]: The skill includes functionality to process untrusted codebase content and metadata artifacts, which serves as a surface for indirect prompt injection. The risk is considered negligible due to the read-only capabilities of the associated tools.
Ingestion points: Codebase files and .socraticodecontextartifacts.json contents.
Boundary markers: Not specified.
Capability inventory: Read-only codebase search, dependency graph querying, and metadata inspection.
Sanitization: Not present.

codebase-exploration