© 2026 YOLOX SYSTEM. ALL RIGHTS RESERVED.
Gives your agent the ability to verify all work with fresh evidence before claiming any task is complete.
When you want to ensure a bug fix actually works before committing
When you need to confirm all tests pass after a significant refactor
When you are about to create a pull request and need proof of success
Ask the agent to perform a specific coding task or fix
Tell the agent which verification commands (tests, lint, build) must run
Review the terminal output provided by the agent as evidence
Verify that the exit codes and failure counts confirm success
You
I've updated the auth logic. Can you make sure everything is still working correctly?
Agent
I will not claim success until I verify it. I'm running `npm test` and `npm run lint` now. [Output: 42/42 tests pass, 0 lint errors]. Based on this evidence, the auth update is verified and complete.
Gives your agent the ability to define clear, unambiguous, and AI-ready specifications for schemas, tools, architecture, and more, ensuring a machine-readable project foundation.
Gives your agent the ability to update technical specifications for requirements, constraints, and data contracts optimized for AI consumption.
Gives your agent the ability to search through historical conversation data to retrieve past decisions, patterns, and insights without wasting context window space.
Gives your agent the ability to analyze chatmode or prompt files and recommend the most optimal AI models based on complexity and cost-efficiency.
Gives your agent the ability to transform a technical specification file into a clear, structured GitHub Issue for a complete feature request.
© 2026 YOLOX SYSTEM. ALL RIGHTS RESERVED.