bro-review-spec-plan

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data (specifications and plans) which may contain malicious instructions intended to influence the agent's behavior.
  • Ingestion points: The skill identifies objects for review from file paths, text passed from other skills, attached files, or the user's current message as described in SKILL.md.
  • Boundary markers: The subagent instructions in subagents/reviewer-prompt.md and subagents/security-reviewer-prompt.md lack explicit delimiters or instructions to treat the ingested content as untrusted data to be ignored if it contains commands.
  • Capability inventory: The skill uses tools to read the repository, file system, and check web links. It is explicitly limited to READONLY operations, which limits the potential impact of an injection.
  • Sanitization: There is no evidence of sanitization or filtering applied to the ingested content before it is processed by the subagents.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 05:08 AM
Security Audit — agent-trust-hub — bro-review-spec-plan