clarity-test-scaffold
Warn
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [METADATA_POISONING]: The skill metadata includes a 'read-only' tag that contradicts its primary purpose of generating and modifying files like vitest configurations and test stubs (SKILL.md).
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from local project files, creating a potential vector for indirect prompt injection.
- Ingestion points: Reads project settings from 'Clarinet.toml' and contract definitions from the 'contracts/' directory (SKILL.md).
- Boundary markers: No delimiters or instructions to ignore embedded instructions are present in the provided skill description.
- Capability inventory: The skill writes new files and updates existing ones, and it executes shell commands using the bun runtime (SKILL.md).
- Sanitization: There is no documentation of input validation or sanitization for the data read from external project files.
- [COMMAND_EXECUTION]: The instructions direct the agent to execute a local script using the bun command-line tool.
- Evidence: 'bun run clarity-test-scaffold/clarity-test-scaffold.ts [options]' in SKILL.md.
- [DYNAMIC_EXECUTION]: The skill relies on the execution of a local TypeScript file ('clarity-test-scaffold.ts') that is not included in the provided analysis set.
Audit Metadata