mission-planning
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE] (SAFE): Comprehensive review across all 10 threat categories shows no indicators of malicious intent or technical risk. The skill acts as a domain-specific knowledge base.
- [PROMPT_INJECTION] (SAFE): Instructions in SKILL.md use standard grounding techniques to ensure the agent follows specific mission planning patterns. There are no attempts to bypass safety filters or extract system prompts.
- [DATA_EXFILTRATION] (SAFE): No network utilities (curl, wget, fetch) or sensitive file path references (SSH, AWS, .env) are present.
- [REMOTE_CODE_EXECUTION] (SAFE): The skill does not include any scripts or package manifests (package.json, requirements.txt) and lacks any remote download or execution patterns.
- [INDIRECT PROMPT INJECTION] (SAFE): While the skill validates user-provided scripts using regex, it lacks any high-privilege capabilities (file-write, network, subprocess) that could be exploited via malicious input data.
Audit Metadata