The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run a local Node.js script located at src/patterns-share.js to perform its core functions of exporting, importing, and displaying file patterns.\n- [DATA_EXFILTRATION]: The skill exports project metadata to a local JSON file. It includes built-in auditing logic designed to prevent data exposure by ensuring only relative paths and frequency counts are included, while explicitly dropping absolute paths or home directory references that could reveal sensitive system information.\n- [PROMPT_INJECTION]: The import command facilitates the ingestion of external data from JSON files, which creates an indirect prompt injection surface.\n
Ingestion points: Data is read from .cco/patterns.digest.json or user-specified paths into the agent's context as described in SKILL.md.\n
Boundary markers: The skill does not implement explicit delimiters or instructions to the agent to treat the imported file content as untrusted data.\n
Capability inventory: The skill is granted Bash and Read capabilities, allowing it to execute local code and access the filesystem.\n
Sanitization: The skill audits imported paths to reject absolute or traversal segments, though it does not describe sanitization for the textual labels or values contained within the JSON file.

cco-patterns