By chrisjenx
Bootstrap a minimal, self-improving /develop orchestration flow fitted to your repo, then grow it via a feedback flywheel.
Code reviewer for /develop:run's PA/PT phases. Reviews the branch diff against the repo's OWN defined conventions and rules (CLAUDE.md, rule docs) plus requirement compliance — does the change do what the spec asked, the way this repo does things. Complements the stack-agnostic general-quality reviewer (fresh-eyes correctness). Reuse-first: applies rules that already exist, doesn't invent standards. Read-only; reports FINDINGS.
Diff-reading completeness auditor for /develop:run's PA phase. Reads the assembled branch diff against the plan's Requirements Inventory and finds requirements that are not actually wired end to end — declared but unused, a layer that should connect but doesn't, a requirement with no corresponding code. Stack-agnostic: reasons about diffs and the requirements table, not about any build tool. Read-only; reports FINDINGS, never edits.
The single generalist executor for /develop:run. Takes a per-phase brief and does the work for exactly that slice of the plan — edits files, runs cheap gates, writes status and findings back to the plan file. Dispatched once per phase by the orchestrator. Stack-agnostic; all repo specifics arrive in the brief.
Fresh-eyes general-quality reviewer for /develop:run's PA/PT phases. Reads the branch diff with no prior context and catches "compiles but obviously wrong" defects — inverted conditions, swapped/wrong arguments, off-by-one, copy-paste bugs, swallowed errors, unclosed resources, obvious security smells. Stack-agnostic: reasons about code logic, not a build tool. Read-only; reports FINDINGS, never edits.
The planner for /develop:run. Reads the spec + repo and emits the plan's Requirements Inventory and Execution Strategy (domain phases with nodes, gate tokens, dependencies, and the existing agent each node routes to). Runs a reuse survey first — reuse existing code, and route to already-defined agents/skills/rules; flag capability gaps to defer, never invent. Dispatched once by the orchestrator so heavy planning context stays out of the loop. Read-only; emits a structured plan.
Manually evaluate the develop flow's accumulated postmortems and decide what to tweak before the next run. Trigger on "/develop:flywheel", "tune the develop flow", "review the postmortems", "what should I improve before the next run", "grow the flow". Reads .claude/develop-flywheel.md across runs, flags recurring/breaking finding categories, and proposes the cheapest remediation lever + target for each — human-gated. Run periodically between feature runs, not per feature (PF already logs each run).
Bootstrap a fitted /develop orchestration flow for THIS repo. Trigger on "init develop", "set up orchestration", "build my own develop flow", "scaffold an agentic pipeline", or any intent to stand up a spec-to-branch loop. Detects the stack, discovers the repo's real build/test/lint gates, and writes the per-repo definitions that the static /develop:run loop reads — it does NOT transplant another project's setup.
Run the fitted /develop orchestration loop for THIS repo — turn a spec, ticket, or rough idea into a reviewed, committed branch. Trigger on "/develop:run", "develop this", "run the develop flow", "spec to branch", or handing this skill a spec/ticket/description. Requires /develop:init to have written .claude/develop.config.json. Walks a per-feature plan file, dispatches one executor per phase, and runs a fixed quality tail wired to the repo's real gates. Static loop; behaviour is fitted because it reads the repo's discovered definitions.
Executes bash commands
Hook triggers when Bash tool is used
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.
Build a self-improving, agentic /develop flow — fitted to your repo, not transplanted from someone else's.
This repo is both a Claude Code plugin marketplace and the home of the develop plugin, plus an interactive explainer of how the underlying orchestration works.
develop — bootstraps a minimal orchestration flow and grows it via a feedback flywheel.Status: early (v0). The bootstrapper skill guides you through the setup today; the deeper automation is tracked in issues.
An agentic /develop pipeline looks like a pile of files, but ~80% of its value is bound to one repo — the gate commands (./gradlew test, npm test, cargo test…), the specialist registry, the conventions. Copying another project's setup installs assumptions that misfire.
So the plugin doesn't copy a finished system. It discovers your repo's real commands, generates a minimal orchestrator (one generalist executor, a plan file, your actual gates), and gives you a flywheel that grows specialists and parallel verification only where repeated pain shows.
The portable part is a small set of moves — state in a file, narrow context per agent, structural gates, close the loop. Everything else is fitted to you.
/plugin marketplace add chrisjenx/cjs-orchestrator
/plugin install develop@cjs-orchestrator
Then, in the repo you want to set up:
/develop:init # once — detect your stack, confirm gates, scaffold the config
/develop:run # repeatedly — spec → reviewed, committed branch
/develop:flywheel # periodically — review postmortems, tune the flow for next run
/develop:init detects your stack, confirms the real gate commands with you, and writes
the repo-specific definitions (gates, conventions, routing, safe hooks) into .claude/.
/develop:run is the portable orchestrator loop: it ships static in the plugin but
behaves fitted to your repo because it reads those discovered definitions. See
DECISIONS.md for the locked names and architecture.
cjs-orchestrator/
├── .claude-plugin/marketplace.json # marketplace catalog
├── plugins/develop/ # the plugin
│ ├── .claude-plugin/plugin.json
│ ├── skills/init/SKILL.md # the /develop:init bootstrapper
│ ├── skills/run/SKILL.md # the /develop:run orchestrator loop
│ ├── skills/flywheel/SKILL.md # the /develop:flywheel tuner
│ ├── references/ # portable mechanism (plan anatomy, briefs, gates)
│ ├── agents/ # portable, stack-agnostic auditor agents
│ └── hooks/ # safe, stack-agnostic hooks
└── docs/ # the explainer site (GitHub Pages)
Changes are in CHANGELOG.md; the release flow + SemVer version policy are in
RELEASING.md. The plugin version (plugins/develop/.claude-plugin/plugin.json)
is what Claude Code uses to surface updates — it's always set, never omitted.
MIT © Chris Jenkins. See LICENSE.
Pattern lineage cross-references Anthropic — Building Effective Agents.
npx claudepluginhub chrisjenx/cjs-orchestrator --plugin developHarness-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
Intelligent draw.io diagramming plugin with AI-powered diagram generation, multi-platform embedding (GitHub, Confluence, Azure DevOps, Notion, Teams, Harness), conditional formatting, live data binding, and MCP server integration for programmatic diagram creation and management.
Feature development with code-architect/explorer/reviewer agents, CLAUDE.md audit and session learnings, and Agent Skills creation with eval benchmarking from Anthropic.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Reliable automation, in-depth debugging, and performance analysis in Chrome using Chrome DevTools and Puppeteer
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review