team-combat
Pass
Audited by Gen Agent Trust Hub on Jul 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or behaviors were detected in the skill's logic or orchestration flow.
- [PROMPT_INJECTION]: The skill processes user-supplied feature descriptions and reads local configuration files (production/review-mode.txt, .agents/docs/technical-preferences.md), which creates a surface for indirect prompt injection. Ingestion points: combat feature description argument and local configuration files. Boundary markers: The skill does not explicitly define delimiters for external data. Capability inventory: The skill has access to tools including Bash, Task, Write, and Edit via the orchestrator or spawned subagents. Sanitization: No explicit sanitization or filtering of input data is performed before processing.
Audit Metadata