By raphavr
Fulcrum — Enterprise software delivery plugin: TDD, systematic debugging, spec-driven development, company standards, metrics collection, and compound learning
Capture and apply learnings from completed development cycles
Execute plan in batches with review checkpoints
Create a structured task specification through codebase analysis and iterative refinement, then generate an implementation plan
Create detailed implementation plan with bite-sized tasks
Use when the user invokes /compound, says "capture learnings", "what did we learn", or when suggesting post-cycle knowledge capture after completing a development branch
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
Use when starting a session, generating specs or plans, reviewing code, or when the user asks about company coding standards, architecture patterns, security requirements, or testing conventions
Use when you have a written implementation plan to execute in a separate session with review checkpoints
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup
Uses power tools
Uses Bash, Write, or Edit tools
No model invocation
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.
Executes directly as bash, bypassing the AI model
Executes directly as bash, bypassing the AI model
A Claude Code plugin for enterprise software delivery teams. Combines Superpowers core skills with enterprise extensions for structured workflows, company standards, metrics, and continuous learning.
Every feature follows the same four-stage flow:
/spec → /write-plan → /execute-plan → /compound
│ │ │ │
Spec Plan doc Working Learnings
artifact saved to branch saved to
saved to docs/plans/ merged or docs/learnings/
docs/specs/ PR open + CLAUDE.md
Each command hands off to the next. You do not run them in isolation.
Fulcrum is not listed in the Anthropic marketplace. Install it by pointing Claude Code at this repo as a custom marketplace:
/plugin marketplace add https://github.com/raphavr/fulcrum
/plugin install fulcrum
/quit # restart Claude Code
/help # confirm /spec and /compound appear
Requires: Claude Code CLI v2.1.0+, Python 3.8+ (for metrics hooks)
/specStart every feature here. Adapts to how clear the idea is.
/spec Add email verification to user registration
/spec I want to improve how we handle errors
docs/specs/YYYY-MM-DD-description.md → you review and approveHands off to /write-plan automatically once the spec is approved.
/write-planTurns the approved spec into a bite-sized implementation plan.
CLAUDE.md for test patterns and commit conventionsdocs/plans/YYYY-MM-DD-feature.md with exact file paths, full code, and TDD stepsHands off to /execute-plan (separate session) or runs subagent-driven execution in the current session.
/execute-planExecutes the plan task by task with review checkpoints.
Hands off to /compound after the branch is finished.
/compoundCaptures what was learned from the completed cycle.
/compound
docs/learnings/ and updates CLAUDE.mdStandards are loaded automatically from config/company-standards/:
| File | Contents | Loaded during |
|---|---|---|
architecture.yaml | Backend/API patterns | Spec |
coding-standards.yaml | Style, naming, line length | Spec, Implement |
security.yaml | Secrets, queries, auth | Spec |
testing.yaml | Coverage, test types | Plan |
git-workflow.yaml | Branch, commit, merge | Plan |
Projects can override any setting by creating .claude/project-config.md:
# Example: checkout-service overrides
architecture:
communication_pattern: "sync-http"
testing:
min_coverage: 95
The session-start hook writes .claude/workflow-session.json (session ID, project, plugin version). The stop hook emits a workflow.session.completed event with fields: duration, stages, commits, outcome.
Events are sent to Kafka. Set these environment variables to enable:
| Variable | Description |
|---|---|
FULCRUM_KAFKA_BROKERS | Comma-separated broker list (required) |
FULCRUM_KAFKA_TOPIC | Topic name for all events (required) |
FULCRUM_KAFKA_CLIENT_ID | Producer client ID (default: fulcrum-plugin) |
FULCRUM_KAFKA_SSL | Enable TLS — true or false (default: false) |
FULCRUM_KAFKA_SASL_USERNAME | SASL plain username (optional) |
FULCRUM_KAFKA_SASL_PASSWORD | SASL plain password (optional) |
When the Kafka variables are not set the hooks run silently and skip emission. Install the Kafka client with pip install confluent-kafka (or pip install -r hooks/requirements.txt).
Edit the YAML files in config/company-standards/ to match your organisation. Distribute changes by bumping the plugin version — engineers run /plugin update fulcrum and pick up the new standards on next session.
npx claudepluginhub raphavr/fulcrum --plugin fulcrumGitHub Spec-Kit integration for Specification-Driven Development - define WHAT and HOW before coding
Skills-first specification-driven development framework with 7 agent skills for planning, implementation, review, and shipping. Natural language activation with intelligent agent orchestration. Includes /plan, /implement, /research commands plus managing-specifications, implementing-features, and reviewing-and-shipping skills.
Specification-driven development workflow: specify → plan → tasks → implement
Spec-driven development for big features. When features get too big, plan mode gets too vague—leading to hallucinations during implementation. ShipSpec replaces vague plans with structured PRDs, technical designs, and ordered tasks that keep Claude grounded.
Specification-Driven Development with Process Discipline for Claude Code
Persona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code