nuxt-test-utils-skilld
Warn
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONMETADATA_POISONINGCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends using
npx -y skilld searchto query documentation. This command downloads and executes theskilldpackage from the NPM registry, which is not a verified vendor resource or trusted service. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process content from various local directories (e.g.,
./.skilld/docs/,./.skilld/issues/,./.skilld/releases/). - Ingestion points: Multiple markdown files within the
.skilld/subdirectory tree. - Boundary markers: Absent; the instructions do not provide delimiters to separate retrieved data from instructions.
- Capability inventory: The skill includes instructions for subprocess execution via
npx. - Sanitization: Absent; external content is processed and searched without explicit validation.
- [METADATA_POISONING]: The skill's metadata contains a future timestamp (
2026-03-19) and references a non-existent model version (Claude Code · Haiku 4.5), representing deceptive/hallucinated metadata.
Audit Metadata