copilot-agents-dojo

Flags risky shell commands and unsafe tree ops.

Detects high-confidence security risks in code.

Surfaces the dojo's non-negotiable prime directives.

Activates the dojo framework at the start of a session.

Plans multi-step work before writing code.

Captures lessons and promotes recurring patterns.

Derives measurable NFRs from a parent business driver.

Derives security requirements from a named threat.

Replace with the capability this skill provides.

Turns vague intent into testable requirements.

Enforces parent-link lineage across requirement layers.

Authors an MCP server with the official SDK and gates.

Bypasses a flaky MCP broker by spawning the server directly.

Wires MCP servers into a Copilot client and verifies them.

Authors new SKILL.md files that conform to the dojo spec.

Reproduces, diagnoses, fixes, and verifies bugs unaided.

Refines rough ideas into approved designs before code.

Reviews diffs by severity to produce actionable feedback.

Maps an unfamiliar repo before touching its code.

Systematic root-cause investigation for hard bugs.

Challenges hacky fixes on non-trivial changes.

Runs independent subtasks concurrently via sub-agents.

Picks the board over sub-agents for cross-turn work.

Executes approved plans one task at a time, verified.

Verifies, summarises, and closes a development branch.

Prepares branches and PRs for clean, reviewable merges.

Processes review feedback until the change is approved.

Safe, test-backed code restructuring in small steps.

Self-reviews work against the plan before sign-off.

Delegates research and parallel work to sub-agents.

Writes meaningful tests that actually catch bugs.

Isolates each task in its own git worktree off main.

Proves work with tests, diffs, and logs before sign-off.