The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external source code files, creating a surface where malicious instructions embedded in the analyzed code could influence the agent's output.
Ingestion points: The Python scripts scripts/complexity_analyzer.py and scripts/dependency_graph.py read Kotlin (.kt) files from the local filesystem.
Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore embedded prompts within the files being analyzed.
Capability inventory: The skill allows the agent to read file contents, extract structural data via regex, and present this information in a summary report.
Sanitization: The scripts use static regex analysis rather than dynamic execution, though they do not sanitize text extracted from comments or identifiers which could contain injection payloads.
[COMMAND_EXECUTION]: The skill includes Python scripts meant to be executed by the agent to perform filesystem operations.
Evidence: The scripts scripts/complexity_analyzer.py and scripts/dependency_graph.py perform directory traversal and file reading. They implement path validation logic (safe_user_path) to prevent path traversal attacks by ensuring that file access remains within the intended project directory.

matematico-tao