skills/romainsimon/paperasse/syndic/Gen Agent Trust Hub

syndic

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill's primary function is property management following French law (Loi 1965, ALUR, ELAN). No malicious instructions or hidden behaviors were found.
  • [SAFE]: Secret management instructions correctly recommend using a .env file for Qonto API keys, which is a secure practice for managing credentials.
  • [COMMAND_EXECUTION]: The skill instructions include executing a Node.js script (integrations/qonto/fetch.js) to fetch bank transactions. This is a legitimate functional requirement for the Qonto banking integration.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted property data from user-defined copros/*.json files.
  • Ingestion points: Property configuration data in copros/*.json files.
  • Boundary markers: None explicitly defined for data interpolation in the provided markdown templates.
  • Capability inventory: Execution of a local Node.js script and writing to local files (.env, copros/*.json, journal/*.md).
  • Sanitization: No explicit sanitization or validation logic was found for the external JSON data inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 09:50 AM
Security Audit — agent-trust-hub — syndic