cs-feat-design

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local commands such as grep to search for concepts in the codebase and python .codestable/tools/validate-yaml.py to check the syntax of generated checklists. These commands are integral to its function as a design tool.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted content from user intents and project source files.
  • Ingestion points: Project documentation in .codestable/, source code, and user-provided {slug}-intent.md files.
  • Boundary markers: None identified; external content is read into the context without explicit delimiters to ignore embedded instructions.
  • Capability inventory: File system write access (design and checklist documents), subprocess calls (grep, validate-yaml.py), and integration with other agent skills.
  • Sanitization: Use of a YAML validator for output files helps ensure structural integrity but does not sanitize instructional content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 12:39 PM
Security Audit — agent-trust-hub — cs-feat-design