mundi-orch-multi-llm-consensus

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly lists reading ~/Mundi Princeps/config/api_keys.json (Anthropic/OpenAI/Gemini keys) and describes invoking providers via curl/Codex/CLI, which requires the agent to access secrets and risks those values being embedded or emitted verbatim in generated requests/commands.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required workflow fans out the user-supplied question to multiple external LLM providers and then feeds the providers’ free-form responses (outsider-authored text from Claude/Gemini/GPT) into the “Aggregator scribe” LLM context for consensus generation.