feature-builder-acrazie
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill suggests installing additional capabilities via
npx skills add Acrazie/skills@.... These are identified as vendor-owned resources under theacrazieaccount and the instructions explicitly require user approval before execution. - [COMMAND_EXECUTION]: The skill is designed to run local test suites and build processes (e.g., 'Run targeted tests', 'existing test runner') to verify feature implementation.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and acts upon external data, including repository instructions, task documents, and approved contracts, which could contain malicious instructions.
- Ingestion points: The skill reads repository instructions, manifests, Git state, and task contracts from the local environment.
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore' directives when processing the content of these files.
- Capability inventory: The agent has the authority to edit files and execute shell commands for testing, creating a high-impact path if injection occurs.
- Sanitization: No explicit sanitization, filtering, or validation of the content within ingested files is described in the workflow.
Audit Metadata