doc-writing
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an ingestion surface for potentially untrusted data.
- Ingestion points: The skill is designed to process source code, issue trackers, dialogue records, and user-provided lists to generate or translate documentation (mentioned in SKILL.md).
- Boundary markers: No explicit instructions are provided for the agent to utilize delimiters or warnings to isolate external content from its primary instructions.
- Capability inventory: The skill primarily performs text writing and translation. While it mentions the use of other tools for API reference generation, it does not define specific subprocess calls, network-write operations, or file-write capabilities within its own instructions.
- Sanitization: No specific input validation, escaping, or filtering logic is described for the processed external data.
- [EXTERNAL_DOWNLOADS]: The skill references a wide range of official documentation and research resources from trusted organizations and well-known services.
- Evidence: Links to repositories from Google (Go), Rust, Django, Curl, SQLite, and Ant Design are provided in references/human-samples.md for stylistic reference.
- Evidence: Informational links to Wikipedia and AI research blogs are included in references/ai-patterns.md.
Audit Metadata