hyfdev-adversarial-review

Pass

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes untrusted repository content, creating a surface where malicious instructions in code or metadata could target subagents. Ingestion points: Git diffs, file reads, and PR metadata (SKILL.md). Boundary markers: A strict briefing template scopes the context provided to reviewers (SKILL.md). Capability inventory: Subagents run verification checks and the main agent automatically applies fixes (SKILL.md). Sanitization: The skill employs fresh subagent instances with no inherited session context to ensure isolation.
  • [DYNAMIC_EXECUTION]: The skill enables high-autonomy behaviors including automated file modification to apply fixes and the execution of repository-specific checks by subagents. These features are standard components for the skill's primary purpose as an automated developer review tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 27, 2026, 06:28 AM
Security Audit — agent-trust-hub — hyfdev-adversarial-review