track-management
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- PROMPT_INJECTION (SAFE): No attempts to override system prompts or bypass safety filters were detected. The instructions are purely functional and focused on the stated task of track management.
- DATA_EXFILTRATION (SAFE): There are no network operations (curl, wget, fetch) or hardcoded credentials. The skill only references local documentation files.
- REMOTE_CODE_EXECUTION (SAFE): No remote scripts or external package installations are defined. The skill consists entirely of markdown instructions.
- INDIRECT_PROMPT_INJECTION (LOW): The skill is designed to process external files like
spec.mdandplan.md. While these represent an ingestion surface for untrusted data, the skill lacks any executable capabilities (scripts, shell access, or network calls) that could be exploited through such an injection.
Audit Metadata