RedBookSkills

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill repeatedly instructs the agent to include an xsec-token (XSEC_TOKEN) and similar sensitive tokens as explicit command-line arguments (e.g., --xsec-token XSEC_TOKEN / feed_id + xsecToken), which requires the LLM to handle and potentially emit secret values verbatim.

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and scrapes public Xiaohongshu pages and arbitrary webpage URLs (see SKILL.md step 5 and the cdp_publish.py functions search_feeds, list_feeds, get_feed_detail and the docs/claude-code-integration.md WebFetch flow) and then uses that extracted, user-generated content to generate drafts and drive actions (publish/comments), so untrusted third‑party content can materially influence agent behavior.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly fetches and injects user-provided webpage content at runtime (see the integration docs/example that fetches a webpage such as https://example.com/article), so arbitrary external URLs supplied at runtime can be loaded and their content fed into the agent's context/draft-generation, which directly influences prompts and outputs.