walden-testability
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the
waldenCLI tool. These includewalden versionfor environment verification,walden statusto fetch project state in JSON format, andwalden validateto ensure design integrity. These commands are project-specific and used to maintain the development workflow. - [EXTERNAL_DOWNLOADS]: The skill references the author's official GitHub repository (
github.com/andrearaponi/walden) as the source for the required CLI tool. It explicitly states that it will not install the CLI itself, requiring the user to have it pre-installed, which prevents unauthorized remote code execution during the setup process. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local project files, presenting an indirect prompt injection surface.
- Ingestion points: Project requirements, design documents, and the
.walden/constitution.mdfile are read into the agent's context. - Boundary markers: No explicit delimiters are mentioned for the content ingestion phase.
- Capability inventory: The skill has the capability to write back to design files and execute shell commands via the
waldenCLI. - Sanitization: The skill relies on the
walden validatecommand to verify that any generated or modified content adheres to the project's schema and standards.
Audit Metadata