By wp-media
Standalone utility agents for Claude Code — changelog, PR description, PHPUnit tests, retrospective, and atomic commits.
Generates a PO-ready grouped changelog from merged PRs since the last release. Reads project config from cadenza.json. Invoked by the /cadenza:changelog skill. Returns the output file path.
Standalone PR description generator. Analyzes the current branch against the base branch, reads changed files and tests, and produces a comprehensive PR description using the project template. Does not push or create the PR — for pipeline use, see release-agent. Invoked by the /cadenza:pr skill.
Analyses completed pipeline runs found in {temp_root}/issues/ and produces a sprint retrospective report: DOD pass rates, loop-back counts, escalation reasons, cycle time by effort size, and recurring failure modes. Suggests concrete AGENTS.md Section 13 entries. Invoked by the /cadenza:retrospective skill.
Writes PHPUnit unit and integration tests for PHP source files. Discovers test paths, naming conventions, and group annotations from the architecture skill and existing tests. Invoked by the /cadenza:test skill. May also be invoked manually for coverage gaps.
Show the Cadenza command map — a quick reference to all available agents and commands.
Generate a PO-ready grouped changelog from merged PRs since the last release.
Analyse changes, create atomic commits with generated messages, then offer to push. Accepts inline hints — "without <file>", "excluding <path>", "only <path>".
Generate a PR description for the current branch using commits, diffs, and the project template.
Analyse completed pipeline runs and surface DOD pass rates, loop-backs, and failure patterns.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
██████╗ █████╗ ██████╗ ███████╗███╗ ██╗███████╗ █████╗
██╔════╝██╔══██╗██╔══██╗██╔════╝████╗ ██║╚══███╔╝██╔══██╗
██║ ███████║██║ ██║█████╗ ██╔██╗ ██║ ███╔╝ ███████║
██║ ██╔══██║██║ ██║██╔══╝ ██║╚██╗██║ ███╔╝ ██╔══██║
╚██████╗██║ ██║██████╔╝███████╗██║ ╚████║███████╗██║ ██║
╚═════╝╚═╝ ╚═╝╚═════╝ ╚══════╝╚═╝ ╚═══╝╚══════╝╚═╝ ╚═╝
Standalone solos. No orchestra required.
Utility agents that each have one job — and do it well, with no pipeline, no orchestration, and no ceremony.
In a concerto, the cadenza is the moment the orchestra pauses and a single instrument plays alone — free, expressive, complete on its own terms.
Cadenza is a Claude Code plugin built on that same idea. It contains specialist agents, each handling one task: generating changelogs, writing PR descriptions, authoring PHPUnit tests, and producing retrospective reports. No pipeline. No spec. No twelve-agent handoff chain. Just ask, and it's done.
Use Cadenza on its own for projects that don't need a full delivery pipeline. Or pair it with Maestro — both plugins share the same config schema and coexist without conflict.
/plugin install cadenza@wp-media
Then add a config file to your project:
// .claude/cadenza.json
{
"ai": {
"repo": "my-org/my-plugin",
"temp_root": ".cadenza",
"display_name": "My Plugin",
"architecture_skill": "my-plugin-architecture"
}
}
That's it. All agents are ready.
Each one performs alone.
| Agent | Command | What it does |
|---|---|---|
changelog-agent | /cadenza:changelog | Collects merged PRs since the last release, groups them by user impact, and writes a PO-ready changelog draft |
pr-agent | /cadenza:pr | Reads your branch commits and diffs, then writes a structured PR description — without pushing anything |
test-writer | /cadenza:test | Discovers your project's test conventions, then authors PHPUnit unit and integration tests for PHP source files |
retrospective-agent | /cadenza:retrospective | Scans completed pipeline runs, surfaces DOD pass rates and loop-back patterns, and suggests concrete AGENTS.md learnings |
| (built-in) | /cadenza:commit | Analyses changed files, groups them into atomic commits with generated messages, then offers to push |
| Command | What it does |
|---|---|
/cadenza | Show the full command map |
/cadenza:changelog [v1.2.3] | Generate a PO-ready changelog from merged PRs |
/cadenza:pr | Write a PR description for the current branch |
/cadenza:test [path/to/File.php] | Author PHPUnit tests for a source file (or auto-detect from branch diff) |
/cadenza:retrospective [date-from date-to] | Analyse completed pipeline runs and surface learnings |
/cadenza:commit [without <file>] | Atomic commits with generated messages, then offer to push |
Every agent reads .claude/cadenza.json — a single file committed in your repo. The agents live in the plugin. The identity lives in the config.
Cadenza plugin (installed once) Your projects
────────────────────────────── ────────────────────────────────────
agents/ project-a/
changelog-agent.md ──────────────► .claude/cadenza.json ← identity
pr-agent.md ──────────────►
test-writer.md ──────────────► project-b/
retrospective-agent.md ───────────► .claude/cadenza.json ← identity
skills/
changelog/ project-c/
pr/ .claude/cadenza.json ← identity
test/
retrospective/
commit/
When Cadenza ships a new version, every project picks it up on the next Claude session — no action needed.
| Key | Required | Description |
|---|---|---|
ai.repo | Yes | GitHub repository in owner/repo format — used for PR and issue links |
ai.temp_root | Yes | Directory where agents write output files (changelogs, PR drafts, reports) |
ai.display_name | Yes | Human-readable project name used in generated output |
ai.architecture_skill | No | Name of the project-specific architecture skill — loaded by test-writer to discover test conventions |
npx claudepluginhub wp-media/cadenzaReal-time agent observer dashboard for Claude Code. Visualizes every tool call, agent spawn, and session lifecycle at zero token cost.
AI delivery pipeline for wp-media WordPress plugins — agents, skills, and orchestration.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Harness-native ECC operator layer - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, selective install profiles, and production-ready workflows for Claude Code, Codex, OpenCode, Cursor, and related agent harnesses
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.