maintain-verification-skill
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external feature files and project source code, creating a surface for indirect prompt injection.
- Ingestion points: Feature files located in the
features/directory and the project's source code files (SKILL.md, Step 2). - Boundary markers: There are no explicit boundary markers or instructions to the subagents to ignore potential instructions embedded within the data they are analyzing.
- Capability inventory: The skill possesses the ability to read and write files, as well as execute shell commands defined in the target skill's lifecycle (Launch, Doctor, Drive, Cleanup).
- Sanitization: The instructions do not specify any sanitization or validation of the content read from source files before it is processed by subagents or used to update the map.
- [COMMAND_EXECUTION]: The skill is designed to execute commands and scripts found in the target verification skill's 'Launch', 'Doctor', 'Drive', and 'Cleanup' models. Although this is the primary intended functionality, it involves the execution of code found within the analyzed project environment.
Audit Metadata