sdpm-vibe

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt instructs the agent to extract and "include all data points, numbers, quotes, facts" from source material and write them verbatim into spec files and composer prompts (and to embed full content in run_python/write_text calls and delegation messages), which will cause any secrets present in the provided source to be copied into outputs and tool calls.

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path ingests outsider-authored free text from the user’s provided URLs via CC-native WebFetch (Step 1: “Read everything the user provided: URLs via CC-native WebFetch”), and that fetched page text can be included in specs/brief.md (Step 3) which is then read by composer sub-agents into the LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly fetches user-provided URLs at runtime via the CC-native WebFetch (user-provided URLs) and then ingests that fetched content as the primary source to generate specs/brief.md and outline.md, meaning remote content can directly control the agent's prompts and outputs.