experiment-archive

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: No security issues detected. The skill performs localized state management operations to archive experiment data using provided helpers.
  • [PROMPT_INJECTION]: The skill acts as a data ingestion point for user-provided metrics and learnings which are interpolated into state files. While it lacks explicit boundary markers, the risk of indirect prompt injection is inherent to its function as a logger.
  • Ingestion points: User-provided experimental results and distilled learnings.
  • Boundary markers: None used in written files.
  • Capability inventory: state_read, state_write, state_append used for local filesystem interaction.
  • Sanitization: None observed for user-supplied strings.
  • [COMMAND_EXECUTION]: Uses shell commands to read and write state variables. The commands are limited to local state management and use hardcoded or user-prompted IDs in a safe manner.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 07:59 AM
Security Audit — agent-trust-hub — experiment-archive