advanced-skill-creator
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, creating an attack surface where malicious instructions in the source material could influence the generated skill's behavior.
- Ingestion points: Untrusted source material including repositories, corpora, manuals, and 'idea dumps' are processed to create new skills (identified in
SKILL.mdandreferences/distill.md). - Boundary markers: The instructions do not mandate the use of delimiters or 'ignore' instructions when interpolating content from source materials into generated skills.
- Capability inventory: The workflow includes the creation of
SKILL.mdfiles and executable scripts (referenced inSKILL.mdandreferences/create.md), which could be leveraged to execute shell commands if poisoned. - Sanitization: There are no instructions for sanitizing, filtering, or validating the input data before it is used to synthesize new instructions or scripts.
Audit Metadata