translate-agent-broker-old-to-new-project

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to read exchange.json end-to-end and "preserve any existing variables (e.g. gemini.apiKey)", which requires copying API key values verbatim into the new exchange.json (even if marked "secret"), causing the LLM to handle and output secrets directly.

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow ingests outsider-authored free text by reading the user-provided V1 files end-to-end (notably spec.instructions) and copying that verbatim into the generated V2 broker orchestrator prompt (system.instructions).