tailrocks-root-cause

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns such as prompt injection, data exfiltration, or unauthorized command execution were detected.
  • [CREDENTIALS_UNSAFE]: The skill includes explicit instructions in 'runtime-trust.md' to cite secret locations and types without ever copying their values into logs or outputs.
  • [INDIRECT_PROMPT_INJECTION]: The skill mitigates risks from untrusted data by instructing the agent to treat repository and fetched content as evidence rather than instructions.
  • [COMMAND_EXECUTION]: The skill defines a restricted environment for running diagnostic commands, requiring frozen inputs, disabled network, and scrubbed secrets.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 03:11 PM
Security Audit — agent-trust-hub — tailrocks-root-cause