assistant-mcp

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt shows and instructs embedding service account tokens/API keys directly into config files and inline environment assignments (e.g., GRAFANA_API_KEY = glsa_xxxx and GRAFANA_API_KEY=glsa_xxxx mcp-grafana...), which encourages the agent to output secret values verbatim.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly directs agents to fetch the discovery endpoint GET https://<GRAFANA_ASSISTANT_HOST>/.well-known/agent.json at runtime to obtain an "Agent Card" describing capabilities, which can directly control agent behavior/prompts during A2A integrations, so this URL is a runtime external dependency that can influence prompts.