writing-for-agents
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains purely educational and instructional content regarding documentation practices for AI agents. It focuses on improving agent predictability and reliability through better writing techniques.
- [NO_CODE]: There are no executable scripts, command-line operations, or network requests present in the provided files. The skill consists entirely of Markdown and YAML configuration.
- [METADATA_POISONING]: The YAML frontmatter in SKILL.md and the configuration in agents/openai.yaml accurately describe the skill's purpose and do not contain deceptive instructions or injection attempts.
- [INDIRECT_PROMPT_INJECTION]: While the skill provides instructions on how to structure agent prompts, it does not ingest untrusted external data or provide a surface for third-party injection attacks.
Audit Metadata