tech-article-writer
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface. \n
- Ingestion points: It uses
WebSearchandWebFetchto bring external, untrusted content into the agent's context (SKILL.md). \n - Boundary markers: There are no instructions or delimiters provided to the agent to isolate fetched data from its core logic or to ignore potential instructions embedded in that data. \n
- Capability inventory: The agent is granted the
Writetool, allowing it to save content to the filesystem. \n - Sanitization: The instructions do not define any validation, filtering, or escaping protocols for the content retrieved from external sources.\n- [NO_CODE]: This skill is entirely composed of markdown instructions and templates. It does not include any executable scripts, binaries, or references to external software packages.\n- [SAFE]: The content was analyzed for direct prompt injection, credential exposure, and obfuscation. No such patterns were detected in the prompt text.
Audit Metadata