anycable-coder

Pass

Audited by Gen Agent Trust Hub on Oct 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides integration patterns for AnyCable, utilizing its official libraries and configuration standards. It references well-known packages (@anycable/web, anycable-rails) and describes standard infrastructure components (AnyCable-Go, Redis, NATS).
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by demonstrating how to pass client-side messages to an LLM through WebSocket channels.
  • Ingestion points: The AssistantChannel and LlmStreamChannel classes in SKILL.md ingest user-provided 'message' and 'prompt' data.
  • Boundary markers: The provided code snippets do not include explicit separators or guardrail instructions for prompt interpolation.
  • Capability inventory: The skill's logic is focused on real-time messaging and message persistence; it does not expose shell access, sensitive file system operations, or unauthorized network capabilities to the processed data.
  • Sanitization: Input validation or sanitization logic is not included in the architectural templates provided.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 9, 2026, 12:28 AM