skills/mrsknetwork/supernova/audit/Gen Agent Trust Hub

audit

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it processes external, untrusted code provided by users.
  • Ingestion points: User-provided code snippets, PR diffs, and module source files as specified in Step 1 of the auditing SOP.
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate untrusted code from the agent's instructions.
  • Capability inventory: The agent is instructed to perform shell-based operations including searching with grep and running analysis tools.
  • Sanitization: No sanitization or escaping of input code is performed prior to the agent's analysis.
  • [COMMAND_EXECUTION]: The skill directs the agent to execute various shell-based commands to perform auditing and static analysis tasks.
  • Evidence: Step 3 and Step 4 provide specific commands such as grep, ruff check, pytest --cov, pip-audit, and npm audit to be run on the target codebase.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 11:19 AM
Security Audit — agent-trust-hub — audit