fast-io

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt explicitly references API keys, Bearer tokens, user passwords, and short-lived org-transfer tokens and tells the agent to accept/provide them (e.g., "provide it as a Bearer token" and claim URLs carrying a token), which creates situations where the agent would need to receive and insert secret values verbatim into requests or outputs.

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow includes fetching the full agent guide from https://mcp.fast.io/skill.md (public web content) and ingesting it as markdown text into the agent’s LLM context, which is an outsider-authored source.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly tells agents to fetch the full agent guide at session start from https://mcp.fast.io/skill.md (and related MCP endpoints), a required runtime fetch whose content defines available tools, parameters, and prompts and therefore directly controls agent behavior.