tk-eli5
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions explicitly prohibit adding external assets, fonts, frameworks, or making network requests, which ensures the generated HTML remains isolated.
- [SAFE]: A collision-free path logic is implemented to ensure that the skill does not overwrite existing files without explicit approval.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user topics to create files, establishing a processing surface for untrusted data. \n
- Ingestion points: User-provided topic strings and optional output paths in SKILL.md. \n
- Boundary markers: The instructions include explicit 'NotApplicable' and 'Unverifiable' conditions to handle out-of-scope or ambiguous requests. \n
- Capability inventory: Writing a single, self-contained HTML file to the local directory. \n
- Sanitization: Filenames are derived from a 'topic-slug', which provides a layer of path sanitization for user input.
Audit Metadata