skills/mblode/agent-skills/ax-audit/Gen Agent Trust Hub

ax-audit

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use shell tools like ripgrep (rg) to search for specific code patterns (e.g., tool execution loops, approval gates, and state machines) across the audited project's source code. These commands are provided as explicit recipes within the rule files to ensure high-fidelity detection while maintaining a read-only posture on the target codebase.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing untrusted third-party code and documentation. This creates an attack surface for indirect prompt injection, where an attacker could embed deceptive instructions or linter-style suppressions (like ax-audit-ignore) in the audited files to manipulate the agent's findings or verdict. The skill mitigates this by providing a self-check workflow and requiring specific evidence for all findings.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 04:06 AM