agent-auditor

Pass

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill implements an audit process that involves reading and analyzing external SKILL.md files from various plugins, which creates a surface for indirect prompt injection. A malicious file being audited could contain instructions designed to influence or deceive the auditor's judgment.\n
  • Ingestion points: The workflow documentation in SKILL.md (Step 2) describes enumerating and reading the contents of external SKILL.md files.\n
  • Boundary markers: No specific delimiters or isolation instructions are provided for the ingested external content.\n
  • Capability inventory: The skill is designed to work with file system tools and a benchmarking harness (bun) that can execute shell commands.\n
  • Sanitization: The methodology does not include steps for sanitizing or validating the content of audited files before they are processed by the agent.\n- [COMMAND_EXECUTION]: The audit workflow and reference documentation describe the use of shell commands such as ls for file enumeration, wc for word count validation, and bun for executing testing benchmarks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 16, 2026, 03:27 AM