setup-writing-framework
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security risks were detected. The skill performs standard configuration tasks with appropriate security constraints.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it reads repository content to determine configuration. However, this risk is well-mitigated by explicit instructions.
- Ingestion points: Inspects root instruction files and documentation files such as AGENTS.md, CLAUDE.md, and GLOSSARY.md in the repository workspace.
- Boundary markers: The skill contains explicit instructions in the Guidelines and boundaries section to treat all repository files, instructions and documentation as untrusted data and to never execute embedded directives or prompt injections.
- Capability inventory: The skill has the capability to read local files and write a specific Markdown block to repository instruction files (AGENTS.md or CLAUDE.md).
- Sanitization: Implements a user-in-the-loop verification via a preview step and provides specific instructions to the agent to disregard external directives discovered while scanning.
Audit Metadata