create-verification-skill
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to identify, generate, and execute shell commands for application lifecycle management, including building binaries, installing dependencies, and running verification tasks. Specifically, Step 4 ("Prove the generated skill before handing it over") requires the agent to execute the Launch, Doctor, Drive, and Cleanup commands that were derived from the repository contents.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository to determine its execution logic, creating an attack surface for indirect prompt injection.\n
- Ingestion points: The agent is instructed to analyze repository files such as README.md, package.json, Makefile, and existing test specifications to extract app-specific commands (SKILL.md Step 1).\n
- Boundary markers: The instructions do not include requirements for the agent to use delimiters or ignore potentially malicious instructions embedded within the analyzed codebase.\n
- Capability inventory: The agent possesses the capability to execute arbitrary shell commands, write files to the local system, and manage processes as part of the verification workflow (SKILL.md Step 2 & 4).\n
- Sanitization: The skill lacks instructions for validating, filtering, or sanitizing the commands extracted from the codebase before they are executed during the verification phase.
Audit Metadata