The Agent Skills Directory

[SAFE]: The skill content is purely instructional, providing guidelines for TDD, interface design, and refactoring. No executable scripts or suspicious commands are included.- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read project-specific documentation such as domain glossaries and Architecture Decision Records (ADRs) from the codebase. Evidence Chain: 1. Ingestion points: codebase files, domain glossary, and ADRs (SKILL.md). 2. Boundary markers: Not present. 3. Capability inventory: The agent is expected to write code and execute tests. 4. Sanitization: Not present. This ingestion is a necessary part of the software development workflow and is considered safe within the context of the skill's primary purpose.

tdd