The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module in Python to execute git for repository discovery and to run its own internal scripts for inventory and validation. These commands are localized to the repository being analyzed and are essential for generating the requested architecture maps.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted code from the local repository.
Ingestion points: The scripts/zoom_out_inventory.py script reads the content of repository files to identify symbols, imports, and caller relationships.
Boundary markers: The skill includes an output-contract.md that defines evidence labels such as 'Verified', 'Likely', and 'Lead' to help the agent categorize the reliability of the information it finds.
Capability inventory: The skill possesses the ability to read arbitrary files within the repository and execute git via subprocess.
Sanitization: The skill performs heuristic scanning of repository content but does not implement explicit content sanitization.
[DATA_EXFILTRATION]: The zoom_out_inventory.py script is configured to recognize .env files as relevant configuration files to search for symbols. While this enables the agent to read potentially sensitive configuration data, there are no network operations or external endpoints present in the skill to facilitate data exfiltration.

zoom-out