designpowers-agent-team
Warn
Audited by Gen Agent Trust Hub on Jun 15, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The installation instructions direct users to clone code from a repository at "https://github.com/Owl-Listener/designpowers.git". This repository is not associated with a verified organization or the stated author.
- [REMOTE_CODE_EXECUTION]: The skill involves installing the cloned repository as a plugin or extension (using commands like "claude plugin install" or "gemini ext add"), which leads to the execution of unverified code within the agent's context.
- [COMMAND_EXECUTION]: The documentation details the use of a "designpowers" CLI for operations like "build", "review", and "taste", which indicates that the skill functions by executing various shell commands.
- [PROMPT_INJECTION]: The skill exhibits a large attack surface for indirect prompt injection due to its handling of untrusted external content.
- Ingestion points: It processes external website content via URLs, user-pasted code snippets, and existing project files like "DESIGN.md".
- Boundary markers: The skill lacks specified markers or warnings to prevent the agent from obeying instructions hidden within the processed external data.
- Capability inventory: The skill has access to a CLI tool, file system write permissions for configuration and documentation, and network access for auditing tasks.
- Sanitization: No sanitization or validation of the input data (URLs, code, or screenshots) is described, leaving the agent vulnerable to malicious payloads.
Audit Metadata