fmt-csv-to-json
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions suggest downloading a helper script (call.mjs) from the vendor domain fittings.sh if it is not already present.
- [CREDENTIALS_UNSAFE]: The skill requires a cryptocurrency private key (FITTINGS_PRIVATE_KEY) to be provided via environment variables to facilitate payments. The documentation notes that this key remains in plaintext within the agent's execution environment.
- [COMMAND_EXECUTION]: The skill includes call.mjs, a Node.js script intended to be executed locally to perform the handshake and payment process with the remote service, involving file system and network access.
- [PROMPT_INJECTION]: The skill processes external CSV data that is converted to JSON and returned to the agent, creating a surface for indirect prompt injection. 1. Ingestion points: The 'csv' field in the POST request body processed by the external service. 2. Boundary markers: No explicit markers are mentioned to isolate untrusted data from agent instructions. 3. Capability inventory: The skill environment has access to the network (fetch), file system (readFileSync), and sensitive environment variables (FITTINGS_PRIVATE_KEY). 4. Sanitization: No evidence of input validation or output sanitization for the CSV content is provided.
Audit Metadata