docs

Fail

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: HIGHDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: Path traversal detected. The skill file contains a relative path (../../../.agents/workflows/docs.md) designed to escape the current directory context.
  • Evidence: The file SKILL.md contains only the string ../../../.agents/workflows/docs.md.
  • Impact: This pattern attempts to access hidden internal directories (.agents) which typically contain sensitive system workflows, metadata, or instructions. Accessing files outside the sandbox is a significant security risk that can lead to information disclosure or privilege escalation within the agent's environment.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 18, 2026, 11:12 AM
Security Audit — agent-trust-hub — docs