cs-decide
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local Python utility located at
.codestable/tools/search-yaml.pyto query metadata from project documentation. This represents execution of shell commands within the project environment. - [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8) as it processes untrusted data from both user inputs and existing project files.
- Ingestion points: Technical decisions, backgrounds, and alternatives provided by the user, as well as the content of
.codestable/attention.mdand related documentation files. - Boundary markers: The skill lacks explicit markers or instructions to isolate or ignore potentially malicious instructions embedded in the user-provided data or existing project files.
- Capability inventory: The agent is authorized to execute a local search script and write new Markdown files to the
.codestable/compound/directory. - Sanitization: There is no evidence of sanitization for user input before it is interpolated into the search tool's command-line arguments or written into permanent project documentation files.
Audit Metadata