Security & Testing Agent Skills for Microsoft Copilot

A behavioral skill that forces every claim the agent makes to be evidence-gated and honestly labeled.

Select and execute the verification that actually proves the change works, based on what type of change it is.

Extracts and verifies the runnability and code coverage of all Python code blocks inside a Markdown file.

Use for a technical incident, disputed diagnosis, bug fix, or claim that a change works.

Verify a test-program change or conclusion through fresh deterministic evidence and iterative repair.

Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare artifacts, and return VERIFIED…

How to check that a Minecraft plugin actually works on a running Paper/Purpur server using VitaminMCP - connecting a session, reading…

Use when about to run tests/builds for ctsnetwork-event, after editing any file in event/backend or event/frontend, or before reporting…

Use when stakes are high — adversarial fresh-context review of every non-trivial decision.