audit-report

Pass

Audited by Gen Agent Trust Hub on Mar 26, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill facilitates read-only access to audit logs via the harness-mcp-v2 server. It implements standard filtering and pagination logic without any dangerous command execution, persistence mechanisms, or privilege escalation patterns.
  • [SAFE]: No unauthorized data exfiltration or hardcoded credentials were found. The skill processes data internally to generate markdown reports based on static templates provided in the references directory.
  • [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) as it processes logs that may contain untrusted strings.
  • Ingestion points: Untrusted data enters the context through the harness_list tool output as defined in SKILL.md.
  • Boundary markers: None present; the skill instructs the agent to analyze and correlate the raw event data directly.
  • Capability inventory: The skill's capabilities are limited to reading logs and generating formatted text reports.
  • Sanitization: No explicit sanitization or filtering of the audit log content is mentioned, which is common for auditing tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 26, 2026, 10:22 PM
Security Audit — agent-trust-hub — audit-report