writing-shape
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill description contains the word 'exploit' in a manner that could be misinterpreted, though context suggests it refers to 'exploiting' raw materials for writing. More importantly, the use of phrases like 'exploit ： 将原始素材逐段塑形' and '这是 exploit（利用）' mimics the tone of prompt injection directives meant to override standard behavior.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process raw, untrusted markdown files provided by users.
- Ingestion points: Raw material files provided by the user (SKILL.md).
- Boundary markers: None identified. The instructions do not define clear delimiters or include warnings for the agent to ignore instructions embedded within the raw material.
- Capability inventory: File read operations (reading the 'pile') and file write operations (appending to the article file).
- Sanitization: None. The skill instructs the agent to 'extract' and 'rewrite' snippets, but lacks explicit sanitization or escaping mechanisms for data handled from the raw material pile.
Audit Metadata