By adam-badar
Core compound engineering workflows, agents, and review gate orchestration.
Explore problem framing and solution options before planning, with optional agent-team fan-out
Capture a non-trivial solved problem as high-signal institutional knowledge (not documentation slop)
Run a structured debug loop to reproduce, isolate, and resolve issues with optional agent-team fan-out
Deepen an existing plan with targeted research passes and iterative PM feedback loops
Run a nested planning loop for mid-epic scope changes and merge approved deltas back into the epic
Synthesizes current best practices from primary docs and high-signal real-world implementations.
Runs freshness-aware external research for volatile domains and time-sensitive assumptions.
Checks framework and library docs to validate implementation assumptions.
Surfaces prior learnings, incidents, and historical pitfalls relevant to the task.
Researches repository context relevant to planning and review questions.
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.
Reusable templates for bootstrapping projects with compound engineering practices: structured AI-assisted development, decision records, runbooks, and CI workflows.
This repository now also contains a private Claude plugin marketplace so shared workflow behavior can be updated in one place and reused across many repositories.
| Path | Purpose |
|---|---|
templates/CLAUDE.md.base | Base Claude Code instructions (branch workflow, planning, memory, circuit breakers) |
templates/.claude/settings.json | Default Claude Code permissions and deny rules |
templates/.claude/settings.marketplace.example.json | Example project settings for marketplace/plugin enforcement |
templates/.claude/agents/codex-gate-runner.md | Legacy compatibility shim for older repos; direct Codex invocation is now canonical |
templates/.claude/agents/codex-pr-correctness-reviewer.md | Legacy compatibility shim for older repos; direct Codex invocation is now canonical |
templates/.claude/agents/codex-pr-edgecase-reviewer.md | Legacy compatibility shim for older repos; direct Codex invocation is now canonical |
templates/.claude/commands/workflows/brainstorm.md | Discovery command to clarify WHAT/why before planning |
templates/.claude/commands/workflows/research.md | Deep research workflow with iterative PM feedback loops |
templates/.claude/commands/workflows/deepen-plan.md | Plan hardening workflow with targeted research passes |
templates/.claude/commands/workflows/plan-loop.md | PM + reviewer iterative planning loop command |
templates/.claude/commands/workflows/debug.md | Structured repro/root-cause workflow with evidence capture |
templates/.claude/commands/workflows/explain.md | Decision and behavior trace workflow for why/how questions |
templates/.claude/commands/workflows/work.md | Execute approved plans with tracker + gate enforcement |
templates/.claude/commands/workflows/pr-review.md | Mandatory PR review command |
templates/.claude/commands/workflows/epic-delta-loop.md | Nested re-planning loop for mid-epic scope changes |
templates/compound-engineering.local.example.md | Review-agent and external-gate configuration template |
templates/docs/adr/ | Architecture Decision Record template |
templates/docs/runbooks/ | Operational runbook template |
templates/docs/runbooks/configure-codex-browser-mcp.md | Codex browser MCP setup for frontend validation |
templates/docs/runbooks/configure-private-marketplace.md | Shared plugin rollout/update runbook |
templates/docs/solutions/ | Solution documentation structure |
templates/docs/process/ | Workflow definitions and sequence diagrams |
templates/docs/plans/ | Plan lifecycle + real-time execution tracker template |
templates/docs/research/ | Deep research artifact structure and conventions |
templates/docs/knowledge/ | Plan index and knowledge base lifecycle guidance |
templates/docs/reviews/ | Review evidence templates and logging conventions |
templates/scripts/ | Worktree creation/removal helpers |
templates/.github/workflows/ | CI workflow templates |
.claude-plugin/marketplace.json | Private Claude plugin marketplace manifest |
plugins/compound-engineering-core/ | Shared plugin with core workflows and agents |
scripts/validate-marketplace.sh | Validates marketplace and plugin manifests |
scripts/release-plugin.sh | Bumps plugin version in both manifests |
./bootstrap/apply-template.sh .
Template command files are still copied for backwards compatibility. Use plugin-prefixed commands in day-to-day work so shared updates come from the marketplace plugin.
claude plugin marketplace add https://github.com/adam-badar/compound-engineering-system.git
claude plugin install compound-engineering-core@compound-engineering-marketplace --scope user
CLAUDE.md.base to CLAUDE.md and customize project-specific sections.compound-engineering.local.example.md to compound-engineering.local.md and customize reviewers.claude mcp add -s user codex-xhigh -- codex mcp-server -c 'model=\"gpt-5.3-codex\"' -c 'model_reasoning_effort=\"xhigh\"' once per machine.frontend-validate (see templates/docs/runbooks/configure-codex-browser-mcp.md).claude plugin list
# should include: compound-engineering-core@compound-engineering-marketplace
npx claudepluginhub adam-badar/compound-engineering-system --plugin compound-engineering-coreCompound Engineering workflow: PRD-driven sprints, isolated worktrees, hook-enforced safety, automated learning. Skills become /vini-workflow:plan, /vini-workflow:compound, etc.
AI-powered development tools. 29 agents, 24 commands, 18 skills, 1 MCP server for code review, research, design, and workflow automation.
Persistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
v9.44.1 — Patch release for Gemini environment/version detection and qwen auth gating. Run /octo:setup.
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
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review