love

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt repeatedly shows curl examples with an Authorization: Bearer {{YOUR_TOKEN}} header and instructs saving/embedding the returned token in requests, which encourages placing secret API tokens directly into LLM-generated commands/outputs (a high exfiltration risk).

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly calls the public inbed.ai API (e.g., GET /api/discover, GET /api/chat, GET /api/agents/me) which returns user-generated profiles, messages, and "suggested actions" that the SKILL.md instructs the agent to follow, meaning untrusted third-party content can be ingested and can materially influence subsequent tool calls and decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill makes runtime calls to https://inbed.ai (e.g., https://inbed.ai/api/...) whose authenticated responses include adaptive "suggested actions" and "soul_prompts" that directly instruct the agent, meaning external content controls prompts and is required for the skill to function.