project-implementer

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: A comprehensive analysis across all threat categories, including prompt injection, data exfiltration, and obfuscation, revealed no malicious patterns.
  • [COMMAND_EXECUTION]: The skill defines a workflow for running build and test commands locally. These operations are restricted to the implementation of planned software tasks and do not involve suspicious network activity or privilege escalation.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes task plans which are external data; however, the protocol mitigates injection risks by requiring each task to be gated by successful test execution and a final manual code review step.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 06:30 AM
Security Audit — agent-trust-hub — project-implementer