write-like-me
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from both the user's conversation and files within the local repository to generate styled prose.
- Ingestion points: Workflow Step 2 in
SKILL.mddirects the agent to gather source text from the conversation, specific files pointed to by the user, and to search for existing documentation within the repository using tools likegrep. - Boundary markers: The instructions do not define delimiters or provide specific warnings to the model to ignore potential instructions embedded within the source content.
- Capability inventory: The skill possesses the capability to read local file content and generate text output. Workflow Step 5 indicates that the resulting text may be passed directly back to other skill workflows, potentially propagating injections.
- Sanitization: There are no provisions for sanitizing or filtering external content before it is interpolated into the drafting process.
Audit Metadata