The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use standard diagnostic utilities such as lsof, ps, tar, and PlistBuddy to inspect system state, process flags, and artifact contents. These actions are legitimate and necessary for the described debugging methodology.
[EXTERNAL_DOWNLOADS]: The skill uses curl to fetch metadata from well-known package registries (e.g., npm) to verify engine constraints and manifest consistency. These operations target trusted domains and are used for artifact inspection rather than executing arbitrary remote scripts.
[DATA_EXFILTRATION]: No sensitive file paths or unauthorized network transmission patterns were identified. While the skill mentions screen capture for visual debugging, this data is intended for the agent's internal visual analysis rather than external transmission to untrusted endpoints.
[PROMPT_INJECTION]: Instructional content was reviewed for behavior override markers, role-play injections, and safety filter bypasses; no malicious prompt injection patterns were found.

signature-repro