spec-driven-development
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides instructions for a legitimate development workflow and does not contain any obfuscated code, credential theft, or unauthorized network operations.
- [COMMAND_EXECUTION]: The skill uses the 'ls' command to locate specification files within the '.claude/specs/' directory. This is a benign operation necessary for its stated functionality.
- [PROMPT_INJECTION]: The skill defines a workflow where external specification files act as the 'source of truth,' creating a surface for indirect prompt injection.
- Ingestion points: Specification files located in '.claude/specs/' (SKILL.md).
- Boundary markers: Absent; the agent is instructed to follow the spec decisions directly.
- Capability inventory: Listing files, reading implementation files, and updating both code and documentation (SKILL.md).
- Sanitization: Absent; the agent relies on the user-provided or project-provided specification content without additional validation steps.
Audit Metadata