The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted text (user-provided articles and drafts) for review and editing. It utilizes tools like Read, Write, and Edit to manipulate this content. While it lacks explicit structural boundary markers for the untrusted content within its instructions, this is a standard risk for content review skills and is managed by the agent's core safety guidelines.
Ingestion points: SKILL.md (reading articles and preference files), references/r5-coherence.md (reading article content).
Boundary markers: None explicitly defined for the article content.
Capability inventory: Read, Write, Edit, AskUserQuestion, and shell tools (grep, find).
Sanitization: No specific sanitization or filtering of external content is mentioned.
[COMMAND_EXECUTION]: The skill uses standard shell utilities such as grep (in references/r1-english.md) and find/grep (in references/r0-factcheck.md) to locate technical terms, technical documentation, or source code snippets within a project. These are used to verify the factual accuracy of the article being reviewed.
[EXTERNAL_DOWNLOADS]: The skill includes links to github.com/gogoingai/wenqu-skills for documentation and project storage. These resources originate from the skill's official vendor and are documented neutrally as standard project infrastructure.

wenqu-review