The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The helper scripts in the scripts/ directory use subprocess.run() with list-based arguments to interact with git, gh, and linearis. This implementation avoids shell interpolation and protects against command injection from user-provided ticket titles or descriptions.
[EXTERNAL_DOWNLOADS] (SAFE): The skill references reputable third-party tools (GitHub CLI, Linearis, OpenSpec) and provides standard installation procedures for these dependencies.
[DATA_EXFILTRATION] (SAFE): While the workflow involves interacting with API tokens (e.g., ~/.linear_api_token), the analysis confirms that no data is transmitted to non-whitelisted or suspicious domains. Communication is restricted to the legitimate APIs of the tools used.
[PROMPT_INJECTION] (SAFE): No instructions designed to override AI safety filters or exfiltrate system prompts were found in the skill's documentation or markdown files.

triage-workflow