colosseum-copilot

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks the user to paste their Personal Access Token and shows commands that embed that token verbatim (TOKEN_VALUE) into saved config and curl requests, which requires the LLM to handle/output secret values directly.

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md under the “Workflow”, the user’s free-form idea/query (“USER_QUERY”) is sent into the runtime via POST requests to https://copilot.colosseum.com/api/v1/search/projects and /api/v1/search/archives, so outsider-authored text (the user prompt) is ingested to drive retrieval/analysis.