skill-creator
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local validation scripts, specifically
python _shared/validate-skills.pyandnode skills/skill-creator/scripts/validate-skill-library.mjs. These commands are used to ensure that new or modified skills adhere to the catalog's structural and metadata standards. This is expected behavior for a developer-oriented authoring tool. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a processing engine for other
SKILL.mdfiles in the repository. This presents a potential attack surface where instructions embedded in the analyzed data could attempt to influence the agent's behavior during the validation or creation process. - Ingestion points: The validation script
scripts/validate-skill-library.mjsreads and parses the content ofSKILL.mdfiles from the local file system. - Boundary markers: The skill relies on standard YAML frontmatter delimiters (
---) and markdown headings to segment data, but it does not include explicit instructions for the agent to ignore potentially malicious prompts contained within the files it is validating. - Capability inventory: The skill uses file system access (read/write) and execution of local Python/Node.js scripts via the command line.
- Sanitization: The validation logic focuses on structural integrity (YAML parsing and regex-based heading checks) rather than sanitizing the natural language content for adversarial instructions.
Audit Metadata