mistral-webhooks-events

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill implementation patterns include network operations to external, user-defined URLs.\n
  • Step 2 and Step 3 demonstrate the use of fetch to send data to a callback URL provided in a request payload. While this is the intended functionality for webhooks, it represents a pattern that allows network requests to arbitrary external domains.\n- [PROMPT_INJECTION]: The skill exhibits surfaces for indirect prompt injection.\n
  • Ingestion points: The skill defines an API endpoint and a background job queue (Step 2) that ingest untrusted data from request bodies, specifically body.messages and body.callback.\n
  • Boundary markers: The provided code examples do not include delimiters or specific instructions to isolate untrusted user input from the broader application context.\n
  • Capability inventory: The skill patterns involve network operations via fetch and interactions with the Mistral AI API for message completion and streaming.\n
  • Sanitization: The examples do not demonstrate validation of the user-provided callback URL or sanitization of the input messages.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 12:46 AM