agent-flight-recorder

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • Data Exposure & Exfiltration (SAFE): The skill writes log files to a local directory (.agents/flight-recorder/). It contains explicit, clear instructions to exclude secrets, tokens, API keys, and PII, directing the agent to summarize or anonymize instead. No network operations or external data transmissions are defined.
  • File System Operations (SAFE): The skill requires standard file system permissions to create and append to markdown/YAML files within the project root. It demonstrates security awareness by instructing the agent to update .gitignore to prevent these logs from being committed to version control.
  • Indirect Prompt Injection (SAFE): As a logging utility, the skill ingests state and error data from the environment. Risk is mitigated by:
  • Ingestion points: Captures 'situation' and 'signal' from task execution.
  • Boundary markers: Uses strictly fenced YAML blocks for all output.
  • Capability inventory: Limited to local file creation and append.
  • Sanitization: Requires the agent to summarize 'why' and anonymize content, reducing the likelihood of persisting malicious payload from untrusted environment signals.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 03:11 PM