skill-host-compat
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local Node.js script located at
scripts/ci/validate-skill-host-compat.jsto perform its linting and validation functions. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an ingestion surface for indirect prompt injection as its primary purpose is to read and process other skill files (
SKILL.md) within the repository. - Ingestion points: The skill reads files from the
skills/,.agents/skills/, and.cursor/skills/directories. - Boundary markers: The instructions do not define clear boundaries or delimiters to isolate the content of the files being linted from the agent's instruction context.
- Capability inventory: The skill uses
nodeto run a script that performs filesystem reads and parsing of Markdown and YAML content. - Sanitization: The skill relies on the validation logic within the
validate-skill-host-compat.jsscript, but does not specify any sanitization or escaping of ingested content to prevent execution of embedded instructions.
Audit Metadata