proj-review

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external source code for review and automated fixing, creating a surface for indirect prompt injection. Malicious instructions could be embedded within code comments or data structures to manipulate the agent's output or behavior during the review process.
  • Ingestion points: External source code files analyzed during the review steps in SKILL.md.
  • Boundary markers: Absent; there are no specific instructions or delimiters to prevent the agent from following instructions embedded within the reviewed code.
  • Capability inventory: Execution of local build tools (mvn compile) and file modification (Step 5: Auto-fix).
  • Sanitization: Absent; the skill does not specify methods to sanitize or escape content from the code files before processing.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute mvn compile in Step 0 to verify code integrity. While this is standard for Java development workflows, it involves running a local build tool on code that may be untrusted.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 10:08 AM
Security Audit — agent-trust-hub — proj-review