The Agent Skills Directory

[PROMPT_INJECTION]: The skill's primary function involves analyzing local repository content, creating a surface for indirect prompt injection where malicious instructions in analyzed files could influence agent behavior.
Ingestion points: Repository files are read and processed by scripts/repo_scan.py, scripts/project_architect.py, and scripts/architecture_diagram_generator.py.
Boundary markers: Absent. The scripts read and output file content directly for processing and display.
Capability inventory: Subprocess execution of git and ripgrep in scripts/repo_scan.py; file read/write operations and command execution mentioned in various agent personas.
Sanitization: Absent. Content from the analyzed repository is processed without explicit escaping or instruction-ignoring delimiters.
[COMMAND_EXECUTION]: The script scripts/repo_scan.py uses subprocess.run to execute standard development tools including git and ripgrep (rg) to discover repository state and observability patterns. While the binaries and most arguments are hardcoded, the script operates on local workspace data.
[EXTERNAL_DOWNLOADS]: The SKILL.md file provides instructions for installing external skills using a script named install-external-skills.py. This mechanism is designed to fetch and install additional functionality from an external registry to extend the agent's capabilities.

architecture