The Agent Skills Directory

[SAFE]: The skill's primary logic is contained in scripts/job_spec.py, which is a simple Python script that processes command-line arguments and prints a JSON object. It does not perform network calls, file system writes, or use dangerous functions like eval() or exec().
[SAFE]: No credentials, API keys, or sensitive file paths (e.g., .ssh, .env) were found in the skill files.
[INDIRECT_PROMPT_INJECTION]: The skill demonstrates an attack surface for indirect prompt injection because the references/prompt-template.md file instructs the agent to process "workspace context." If the workspace contains files with malicious instructions, the agent could potentially be influenced. However, the prompt template explicitly includes safety guidelines such as "avoid background writes unless the user has opted into them" and "keep the scope explicit," which are defensive practices.
[SAFE]: The skill does not attempt to implement persistence on the host system (like modifying shell profiles or crontabs). It merely generates a specification that the user is expected to integrate into their own automation flow.

kairos-lite