flow-code-review

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious code, exfiltration patterns, or obfuscation detected. The skill instructions focus on standard software development lifecycle (SDLC) verification tasks.
  • [PROMPT_INJECTION]: The skill operates as a reviewer of external data, such as task specifications, PRDs, and implementation code. While this creates a surface for indirect prompt injection, the skill itself does not contain malicious instructions.
  • Ingestion points: Implementation code, PRD documents, Technical Specifications, and Task specs.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are provided for the processed input data.
  • Capability inventory: File writing (producing review results), task rejection/approval logic, and calling other local skills via the stack selector.
  • Sanitization: No explicit sanitization or filtering of external content is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 09:25 PM