klingai-webhook-config

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external webhook events, which is an inherent part of its functionality.
  • Ingestion points: Webhook endpoints defined in SKILL.md and references/implementation.md (e.g., /webhooks/klingai) ingest JSON data from the Kling AI service.
  • Boundary markers: The instructions explicitly mandate validating the X-Kling-Signature header using HMAC SHA-256 to ensure data authenticity.
  • Capability inventory: Code templates provide hooks for downloading videos, updating databases, and notifying users, which are common event-driven actions.
  • Sanitization: The skill includes instructions to enforce schema validation and least-privilege routing to mitigate the risk of malicious payload content.
  • [CREDENTIALS_UNSAFE]: The skill demonstrates safe secret management by instructing the use of environment variables (KLING_ACCESS_KEY, KLING_SECRET_KEY, KLINGAI_WEBHOOK_SECRET) for API authentication and signature verification, rather than hardcoding secrets.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with the official Kling AI API (api.klingai.com) for task management and webhook registration, which is the intended purpose of the skill.
  • [COMMAND_EXECUTION]: The skill provides standard implementation examples for Flask and Express.js servers intended for the developer to run in their own environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:45 AM
Security Audit — agent-trust-hub — klingai-webhook-config