gitnexus-pr-swarm-review
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill follows a structured orchestration process for code reviews and explicitly prohibits any destructive actions, such as editing or committing code.
- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it processes content from external pull requests.
- Ingestion points: Data is pulled from external repositories via the provided PR URL or number.
- Boundary markers: None explicitly defined in the adapter file; behavior is governed by an external orchestration document.
- Capability inventory: Uses the platform's
Agenttool for task delegation; instructions explicitly mandate a "read-only" investigative mode. - Sanitization: No explicit sanitization or filtering of PR content is performed in the skill instructions.
Audit Metadata