measuring-social-performance

Pass

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE] (SAFE): The skill file consists entirely of Markdown text for agent instructions. There are no scripts, binaries, or configuration files that could execute commands or perform network operations.
  • [INDIRECT_PROMPT_INJECTION] (SAFE): The skill is designed to ingest external data (CSV files and user-provided analytics) but lacks any technical capabilities to act on malicious instructions within that data.
  • Ingestion points: User answers to discovery questions and uploaded CSV analytics files.
  • Boundary markers: None provided in the prompt text.
  • Capability inventory: None. No shell execution, file system writing, or network request functions are defined in this skill.
  • Sanitization: Not applicable as the skill contains no processing logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 18, 2026, 01:06 PM