Browse Skills

Trigger ADO pipelines for a Copilot-created PR by posting /azp run comments. Use when the user asks to trigger CI pipelines for a specific PR.

Establishes or extends a domain-grounded design system by researching real-world prior art, applying taste, and avoiding common design pitfalls. Use when designing a UI, visual system, or making a screen look good.

Automates the hyperflow ship phase: runs pre-push gates (lint, typecheck, build, tests, security sweep), reports results, and then asks for confirmation before pushing. Never bypasses hooks or force-pushes to main.

A memory manager that lets AI agents view, search, add, edit, prune, or clear Hyperflow project memory stored in .hyperflow/memory/, without touching source code.

Performs multi-level code reviews (from syntax checks to adversarial probing) on diffs, commits, branches, or PRs, with mandatory security scanning at higher levels. Findings are written to .hyperflow/audits/ and a fix-gate can route fixes to planning.

Provides a big-task workflow lane for system-wide changes, large migrations, and repo-wide audits. Uses Claude Code's native dynamic workflows or a portable adapter for Codex, OpenCode, and Grok.

Systematic root-cause analysis for bugs, test failures, runtime errors, and broken builds. Use before patching to identify structural causes, not symptoms, with a step-by-step agent workflow.

Sets the auto-routing mode for Hyperflow, controlling when user messages automatically trigger Hyperflow chains. Supports full routing, intent-based routing, or disabling auto-routing.

Displays a one-screen snapshot of the hyperflow project state, including version, profile freshness, memory entries, and live progress on active tasks and features. Useful for quickly checking project status and in-flight work.

One-shot project setup for Hyperflow that analyzes the codebase, builds caches, seeds memory files, and optionally installs detection shims for AI tools like Claude Code and Cursor.

Runs L1-L5 audits against a PR's diff, posts findings as a batched inline or summary review, provides a fix chain for needed changes, and optionally merges approved PRs.

A planning phase for Claude Code that amplifies prompts, designs architectural approaches, and decomposes work into actionable task files without writing implementation code.