fishjam-python-server-sdk

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: Credential Management. The skill correctly instructs developers to use environment variables (e.g., FISHJAM_MANAGEMENT_TOKEN, GEMINI_API_KEY) for sensitive credentials instead of hardcoding them within the application code.
  • [EXTERNAL_DOWNLOADS]: The skill guides users to install the fishjam-server-sdk from PyPI, which is the official library provided by the vendor (software-mansion-labs).
  • [SAFE]: Webhook Implementation. The documentation recommends the use of path-secret patterns to authenticate incoming webhook requests, providing a standard security measure for stateless callbacks.
  • [PROMPT_INJECTION]: Indirect Prompt Injection surface analysis.
  • Ingestion points: Untrusted data enters the context through peer metadata (PeerOptions), webhook bodies, and media streams (audio/images) received via session.receive().
  • Boundary markers: No explicit instructions are provided to the agent to disregard commands embedded within the media or metadata streams, though this is typical for SDK-level documentation.
  • Capability inventory: The skill provides network access capabilities via the FishjamClient and google.genai clients.
  • Sanitization: Structural validation is performed via Protobuf decoding in the receive_binary function, which reduces the risk of malformed data injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 01:40 PM
Security Audit — agent-trust-hub — fishjam-python-server-sdk