claudeers.

🔓 unclaimed — this page was auto-generated from GitHub. Are you the creator?

Claim this page →
// Developer Tools

claude-forge

Supercharge Claude Code with 11 AI agents, 36 commands & 15 skills — the claude-code plugin framework inspired by oh-my-zsh. 6-layer security hooks included.…

// Developer Tools[ cli ][ api ][ web ][ mobile ][ claude ]#claude#agents#ai-assistant#ai-coding#ai-framework#ai-pair-programming#anthropic#devtoolsMIT$open-sourceupdated 15 days ago
Actively maintained
100/100
last commit 9 days ago
last release 9 days ago
releases 4
open issues 13
// install
git clone https://github.com/sangrokjung/claude-forge
Claude Forge

Turn Claude Code into a full development environment

What's New v3.0Quick StartWorkflowsWhat's InsideInstallationArchitectureCustomization한국어

🎉 v3.0.2 released (May 2026) — docs-only patch on top of v3.0.1: LLM-readable install paths (root INSTALL.md + above-the-fold one-liner) and multi-channel distribution. The v3.0.1 baseline brought Anthropic 2026 standard alignment (Hooks 21+ events · Subagent frontmatter v2 · Skills/Commands hybrid policy) plus a 4-server MCP minimum (playwright · context7 · jina-reader · [email protected]). See MIGRATION.md / MIGRATION.ko.md, Release v3.0.2, Release v3.0.1.

🚀 Install in one line (full install, recommended):

curl -fsSL https://raw.githubusercontent.com/sangrokjung/claude-forge/main/install.sh | bash

Or, inside an existing Claude Code session: /plugin marketplace add sangrokjung/claude-forge then /plugin install claude-forge (lightweight, plugin-only). Full details: INSTALL.md.


What is Claude Forge?

Claude Forge is an open-source development environment for Claude Code that provides 11 specialized agents, 33 slash commands, 24 skill workflows, 15 automation hooks (plus 9 opt-in examples covering 21 lifecycle events), and 9 rule files. Often described as "oh-my-zsh for Claude Code", it transforms Claude Code from a basic CLI into a full-featured development environment. One install gives you agents, commands, skills, hooks, and rules — all pre-wired and ready to go.

Think of it as oh-my-zsh for Claude Code: the same way oh-my-zsh enhances your terminal, Claude Forge supercharges your AI coding assistant.


⚡ Quick Start

Option 1 — Claude Code Plugin (partial coverage, v3.0.1+)

Inside Claude Code, register the marketplace once, then install the plugin:

/plugin marketplace add sangrokjung/claude-forge
/plugin install claude-forge

Update later via /plugin update claude-forge (or from the /plugin UI).

⚠️ Partial coverage — please read before choosing. The Claude Code plugin loader currently recognizes commands/ and (most of) skills/ but does not auto-wire agents/, hooks/, rules/, statusLine, settings.json env blocks, or the entries inside mcp-servers.json. That is a loader limitation, not a claude-forge one — see docs/PLUGIN-VS-INSTALL-SH.md for the full matrix. If you want every resource (agents, hooks, rules, MCP, statusLine) wired up, use Option 2 below.

# 1. Clone (submodules optional — only needed for the CC CHIPS status bar)
git clone --recurse-submodules https://github.com/sangrokjung/claude-forge.git
cd claude-forge

# 2. Install (creates symlinks under ~/.claude)
./install.sh

# 3. Launch Claude Code
claude

install.sh symlinks every resource into ~/.claude/, so git pull updates instantly, and it is the only install path that delivers agents, hooks, rules, the 4 MCP servers, and the status bar in one shot. Clone without --recurse-submodules still works: the CC CHIPS submodule is optional and the installer skips it cleanly if the directory is empty (a one-line hint is printed).

Which option should I pick?

ResourceOption 1 (/plugin install)Option 2 (./install.sh)
Commands (33)
Skills (24)⚠️ partial¹
Agents (11)
Hooks (15 + 9 examples)❌²
Rules (9)
MCP servers (4)❌³
statusLine (CC CHIPS)✅ (opt-in submodule)
settings.json env

¹ Plugin-loaded skills are supported, but QJC skill features that reach into ~/.claude/rules/ or ~/.claude/agents/ assume the symlink layout. ² See hooks/hooks.json: Claude Code does not load a separate hooks.json; all hook settings must live in settings.json, which Option 1 does not merge. ³ plugin.json does not currently surface mcpServers in a way the loader acts on. MCP servers come from .mcp.json / mcp-servers.json, which only Option 2 wires up.

Recommendation: Use Option 2 unless you only need Commands + a subset of Skills.

If you find Claude Forge useful, please consider giving it a star -- it helps others discover this project.

🎉 What's New in v3.0

ChangeDescription
Hooks 21 EventsLifecycle hooks expanded from 5 to 21 events. Opt-in samples live at hooks/examples/ with the full catalog in hooks/README.md.
Subagent Frontmatter v210 optional fields: isolation, background, memory, maxTurns, skills, mcpServers, effort, hooks, permissionMode, disallowedTools. Schema: reference/agent-schema.json. Details: docs/AGENT-FRONTMATTER-V2.md.
Skills/Commands Hybrid PolicyClear boundary documented at docs/SKILLS-VS-COMMANDS.md. 8 directory-form commands are being migrated to skills/ with symlink compatibility preserved.
MCP Minimal (4 servers, v3.0.1)Default set: playwright · context7 · jina-reader · [email protected] (Google ChromeDevTools org, Apache-2.0 — promoted in v3.0.1 for Lighthouse / Core Web Vitals audits). The legacy full set lives in mcp-servers.optional.json. Recipes: docs/MCP-MIGRATION.md. Decision rationale: docs/adr/ADR-001-mcp-default-set.md.
CLAUDE.md Template + @importNew setup/CLAUDE.md.template and docs/CLAUDE-MD-GUIDE.md with a 200-line principle and @import pattern for modular project instructions.
settings.json 2026 FieldsNew fields: tui (flicker-free rendering), disableSkillShellExecution (sandbox), enabledMcpjsonServers (explicit allowlist).
Upgrade in One Command./install.sh --upgrade safely migrates existing v2.1 installs with backup and diff preview.

🔧 What's New in v3.0.1 (patch)

ChangeDescription
Plugin Manifest shipped (partial)/plugin marketplace add sangrokjung/claude-forge + /plugin install claude-forge now work for Commands + Skills. .claude-plugin/plugin.json + .claude-plugin/marketplace.json both pinned to 3.0.2; CI enforces version drift via the new marketplace-schema job. Agents / Hooks / Rules / MCP / statusLine still require ./install.sh — see docs/PLUGIN-VS-INSTALL-SH.md.
Chrome DevTools promotedLighthouse / Core Web Vitals / memory snapshots now arrive with the default 4-server MCP set. Pinned at [email protected] (supply-chain hardening).
hooks/_lib/timing.shNew wrapper records SessionEnd hook timing into ~/.claude/logs/hook-timing.jsonl (mode 600) so the real parallelism of async: true hooks can be audited post-hoc. ~35 ms overhead.
CI trigger expanded.github/workflows/validate.yml now runs on every PR and on main/feat/**/fix/**/chore/**/docs/**/ci/** pushes (previously main only). 6 jobs total.
Tier 0 spec correctionsHook types aligned to command/http/prompt/agent (were llm-prompt/mcp-tool). timeout unit corrected to seconds (was ms). Auto Memory path ~/.claude/projects/<project>/memory/ (was <hash>).
New governance docsdocs/adr/ADR-001-mcp-default-set.md (MCP default decision, MADR) · docs/SETTINGS-COMPATIBILITY.md (UNVERIFIED field tracking) · docs/MARKETPLACE-SUBMISSION.md (official directory submission packet).
4-way independent reviewsuper-research (Tier 0 docs) · security-reviewer · architect · codex-reviewer ran in parallel on the patch. 11 blocking issues closed before merge.

🚨 Breaking Changes

  • MCP defaults cutmemory, exa, github, and fetch were removed from mcp-servers.json. Restore any of them from mcp-servers.optional.json if you need them. Built-in replacements: Auto Memory, the gh CLI, WebSearch, and the jina-reader fallback cover most previous use cases.
  • 8 commands moved to skills/ — Symlink compatibility is maintained until 2027-04-01. Affected: debugging-strategies, dependency-upgrade, evaluating-code-models, evaluating-llms-harness, extract-errors, security-compliance, stride-analysis-patterns, summarize.
  • settings.json allowlist — Removed mcp__memory, mcp__exa, mcp__github, mcp__fetch. Added mcp__playwright.

Why v3.0?

The 2026 Anthropic Claude Code standard evolved significantly (Skills/Commands integration, 21 hook events, expanded subagent frontmatter). v3.0 aligns fully with that standard, cuts default MCP dependencies from 6 down to 4 (v3.0 shipped a 3-server minimum; v3.0.1 promoted [email protected] into the default set) to lower the barrier for new users, and incorporates dotclaude operational experience (security allowlists, breakage-safe migration). Existing v2.1 users can migrate with a single line: ./install.sh --upgrade.

New here?

If you're new to development or Claude Code, start with these:

StepWhat to do
1Run /guide after install -- an interactive 3-minute tour
2Read First Steps -- glossary + TOP 6 commands
3Browse Workflow Recipes -- 5 copy-paste scenarios

Or just type /auto login page and let Claude Forge handle the entire plan-to-PR pipeline for you.


🔄 Development Workflows

Feature Development Workflow

Real-world workflows that chain commands, agents, and skills together.

Feature Development

Build new features with a plan-first, test-first approach:

/plan → /tdd → /code-review → /handoff-verify → /commit-push-pr → /sync
graph LR
    P["/plan<br><small>Design & risk analysis</small>"] --> T["/tdd<br><small>Tests first, then code</small>"]
    T --> CR["/code-review<br><small>Quality & security check</small>"]
    CR --> HV["/handoff-verify<br><small>Fresh-context validation</small>"]
    HV --> CPR["/commit-push-pr<br><small>Commit, push, PR & merge</small>"]
    CPR --> S["/sync<br><small>Sync project docs</small>"]

    style P fill:#533483,stroke:#fff,color:#fff
    style T fill:#0f3460,stroke:#fff,color:#fff
    style CR fill:#0f3460,stroke:#fff,color:#fff
    style HV fill:#e94560,stroke:#fff,color:#fff
    style CPR fill:#1a1a2e,stroke:#fff,color:#fff
    style S fill:#16213e,stroke:#fff,color:#fff
StepWhat happens
/planAI creates an implementation plan. Waits for your confirmation before coding.
/tddWrite tests first, then code. One unit of work at a time.
/code-reviewSecurity + quality check on the code you just wrote.
/handoff-verifyAuto-verify build/test/lint all at once.
/commit-push-prCommit, push, create PR, and optionally merge -- all in one.
/syncSync project docs (prompt_plan.md, spec.md, CLAUDE.md, rules).

Bug Fix

Fast turnaround for bug fixes with automatic retry:

/explore → /tdd → /verify-loop → /quick-commit → /sync
StepWhat happens
/exploreNavigate the codebase to find where the bug lives.
/tddWrite a test that reproduces the bug, then fix it.
/verify-loopAuto-retry build/lint/test up to 3 times with auto-fix on failure.
/quick-commitFast commit for simple, well-tested changes.
/syncSync project docs after commit.

Security Audit

Comprehensive security analysis combining CWE and STRIDE:

/security-review → /stride-analysis-patterns → /security-compliance
StepWhat happens
/security-reviewCWE Top 25 vulnerability scan + STRIDE threat modeling.
/stride-analysis-patternsSystematic STRIDE methodology applied to system architecture.
/security-complianceSOC2, ISO27001, GDPR, HIPAA compliance verification.

Team Collaboration

Agent Teams Hub-and-Spoke

Parallel multi-agent execution for complex tasks:

/orchestrate → Agent Teams (parallel work) → /commit-push-pr
StepWhat happens
/orchestrateCreate an Agent Team with file-ownership separation and hub-and-spoke coordination.
Agent TeamsMultiple agents work in parallel on frontend, backend, tests, etc.
/commit-push-prMerge all work, verify, and ship.

Why Claude Forge?

Most developers either use Claude Code with no customization or spend hours assembling individual configs. Claude Forge gives you a production-ready setup in 5 minutes.

FeatureClaude ForgeBasic .claude/ SetupIndividual Plugins
Agents11 pre-configured (frontmatter v2)Manual setup requiredVaries by plugin
Slash Commands33 ready-to-use (hybrid policy)NonePer-plugin basis
Skill Workflows24 multi-step pipelinesNonePer-plugin basis
Hooks15 built-in + 9 opt-in examples (21 events)None by defaultPer-plugin basis
MCP Servers3 minimal (8+ optional)NonePer-plugin basis
Installation5 min, one commandHours of manual configPer-plugin install
Updatesgit pull + ./install.sh --upgradeManual per-filePer-plugin update
Workflow IntegrationEnd-to-end pipelines (plan to PR)Disconnected toolsNot integrated

📦 What's Inside Claude Forge

Claude Forge Components

CategoryCountHighlights
Agents11planner architect code-reviewer security-reviewer tdd-guide database-reviewer + 5 more (frontmatter v2)
Commands33/commit-push-pr /handoff-verify /explore /tdd /plan /orchestrate /security-review ...
Skills24build-system security-pipeline eval-harness team-orchestrator session-wrap ... (+8 migrated from commands/)
Hooks15 + 9 examples15 built-in (secret filtering, remote command guard, DB protection, security auto-trigger, rate limiting ...) + 9 opt-in samples covering 21 lifecycle events
Rules9coding-style security git-workflow golden-principles agents interaction verification ...
MCP Servers4 (minimal)playwright context7 jina-reader [email protected] — 7+ more available in mcp-servers.optional.json

📥 Claude Forge Installation Guide

As a Claude Code Plugin (v3.0.1+, partial coverage)

Since v3.0.1 claude-forge ships a standard Claude Code plugin manifest (.claude-plugin/plugin.json + .claude-plugin/marketplace.json). Install via the standard two-step flow:

/plugin marketplace add sangrokjung/claude-forge
/plugin install claude-forge

Upgrade via /plugin update claude-forge (or the /plugin UI). What this path covers today (per Claude Code's current plugin loader): Commands + most Skills. Agents, Hooks, Rules, the 4 MCP servers, the statusLine, and settings.json env blocks still require the classic ./install.sh below. See docs/PLUGIN-VS-INSTALL-SH.md for the resource-by-resource matrix. Most users should keep ./install.sh as the primary install.

Prerequisites

DependencyVersionCheck
Node.jsv22+node -v
Gitanygit --version
jqany (macOS/Linux)jq --version
Claude Code CLI≥1.0claude --version

macOS / Linux

git clone --recurse-submodules https://github.com/sangrokjung/claude-forge.git
cd claude-forge
./install.sh                     # fresh install
# or
./install.sh --upgrade           # safe v2.1 → v3.0 migration (backup + diff preview)

The installer:

  1. Checks dependencies (node, git, jq)
  2. Initializes git submodules (CC CHIPS status bar)
  3. Backs up existing ~/.claude/ if present
  4. Creates symlinks for 7 directories + settings.json to ~/.claude/
  5. Applies CC CHIPS custom overlay
  6. Optionally installs MCP servers and external skills
  7. Adds shell aliases (ccclaude, ccrclaude --resume)

Because it uses symlinks, git pull in the repo updates everything instantly. Use --upgrade after pulling v3.0 to apply MCP/settings migrations in place.

Windows

# Run PowerShell as Administrator
.\install.ps1

Windows uses file copies instead of symlinks. Re-run install.ps1 after git pull to update.

MCP Server Setup

v3.0.1 ships with 4 minimal defaults. Others are available opt-in via mcp-servers.optional.json. Full recipes: docs/MCP-MIGRATION.md. Rationale: docs/adr/ADR-001-mcp-default-set.md.

ServerDefault?API KeySetup
playwrightNot requiredAuto-installed via install.sh
context7Not requiredAuto-installed via install.sh
jina-readerNot requiredAuto-installed via install.sh
chrome-devtoolsNot requiredAuto-installed via install.sh (pinned @0.23.0, Google ChromeDevTools org, Apache-2.0)
memoryopt-inNot requiredMerge from mcp-servers.optional.json
fetchopt-inNot requiredRequires uvx (Python)
exaopt-inOAuthclaude mcp add exa --url https://mcp.exa.ai/mcp
githubopt-inPATSet GITHUB_PERSONAL_ACCESS_TOKEN env var

Customization

Override settings without modifying tracked files:

cp setup/settings.local.template.json ~/.claude/settings.local.json
vim ~/.claude/settings.local.json

settings.local.json merges on top of settings.json automatically.


🏗 Claude Forge Architecture

Skills vs Commandsskills/ host auto-invocable knowledge and reusable workflows (Claude discovers them via the description trigger). commands/ host explicit side-effect actions that the user times by typing /name. See docs/SKILLS-VS-COMMANDS.md for the policy.

Symlink Architecture

graph TB
    subgraph REPO["claude-forge (git repo)"]
        A["agents/ (11)"]
        C["commands/ (33)"]
        S["skills/ (24)"]
        H["hooks/ (15)"]
        R["rules/ (9)"]
        SC["scripts/"]
        CC["cc-chips/"]
        K["knowledge/"]
        REF["reference/"]
        SET["settings.json"]
    end

    INSTALL["./install.sh"]
    REPO --> INSTALL

    subgraph HOME["~/.claude/ (symlinked)"]
        HA["agents/"]
        HC["commands/"]
        HS["skills/"]
        HH["hooks/"]
        HR["rules/"]
        HSC["scripts/"]
        HCC["cc-chips/"]
        HSET["settings.json"]
    end

    INSTALL -->|symlink| HOME

    CLAUDE["claude (CLI)"]
    HOME --> CLAUDE

    style REPO fill:#1a1a2e,stroke:#e94560,color:#fff
    style HOME fill:#0f3460,stroke:#16213e,color:#fff
    style INSTALL fill:#e94560,stroke:#fff,color:#fff
    style CLAUDE fill:#533483,stroke:#fff,color:#fff
Full Directory Tree
claude-forge/
  ├── agents/                    Agent definitions (11 .md files, frontmatter v2)
  ├── cc-chips/                  Status bar submodule
  ├── cc-chips-custom/           Custom status bar overlay
  ├── commands/                  Slash commands (33 .md, 8 dirs moved to skills/)
  ├── docs/                      Screenshots, diagrams, policy docs (v3.0 guides)
  ├── hooks/                     Event-driven shell scripts (15)
  │   └── examples/              Opt-in .example samples for 21 lifecycle events (9)
  ├── knowledge/                 Knowledge base entries
  ├── reference/                 Reference docs (+ agent-schema.json)
  ├── rules/                     Auto-loaded rule files (9)
  ├── scripts/                   Utility scripts
  ├── setup/                     Installation guides + CLAUDE.md template
  ├── skills/                    Multi-step skill workflows (24, hybrid policy)
  ├── install.sh                 macOS/Linux installer (--upgrade supported)
  ├── install.ps1                Windows installer (--upgrade supported)
  ├── mcp-servers.json           MCP server defaults (4 minimal)
  ├── mcp-servers.optional.json  Optional MCP servers (memory/exa/github/fetch/time/...)
  ├── .claude-plugin/plugin.json Plugin manifest (3.0.2)
  ├── .claude-plugin/marketplace.json  Marketplace entry (3.0.2)
  ├── settings.json              Claude Code settings (2026 fields)
  ├── MIGRATION.md               v2.1 → v3.0 migration guide (EN)
  ├── MIGRATION.ko.md            v2.1 → v3.0 migration guide (KO)
  ├── CONTRIBUTING.md            Contribution guide
  ├── SECURITY.md                Security policy
  └── LICENSE                    MIT License

🛡 Claude Code Automation Hooks

Security Hooks

6-Layer Security Defense

HookTriggerProtects Against
output-secret-filter.shPostToolUseLeaked API keys, tokens, passwords in output
remote-command-guard.shPreToolUse (Bash)Unsafe remote commands (curl pipe, wget pipe)
db-guard.shPreToolUseDestructive SQL (DROP, TRUNCATE, DELETE without WHERE)
security-auto-trigger.shPostToolUse (Edit/Write)Vulnerabilities in code changes
rate-limiter.shPreToolUse (MCP)MCP server abuse / excessive calls
mcp-usage-tracker.shPreToolUse (MCP)Tracks MCP usage for monitoring

Utility Hooks

HookTriggerPurpose
code-quality-reminder.shPostToolUse (Edit/Write)Reminds about immutability, small files, error handling
context-sync-suggest.shSessionStartSuggests syncing docs at session start
session-wrap-suggest.shStopSuggests session wrap-up before ending
work-tracker-prompt.shUserPromptSubmitTracks work for analytics
work-tracker-tool.shPostToolUseTracks tool usage for analytics
work-tracker-stop.shStopFinalizes work tracking data
task-completed.shTaskCompletedNotifies on subagent task completion
expensive-mcp-warning.sh-Warns about costly MCP operations

Opt-in Examples (v3.0)

9 additional .example samples covering the newer lifecycle events (SessionEnd, PreCompact, SubagentStart/Stop, MessageStart/End, UserPromptReceived, and more) ship in hooks/examples/. The full 21-event catalog with recipes is in hooks/README.md. Rename *.example*.sh and register in settings.json to activate.


🤖 Claude Code Agents

Each agent has a color in the UI for quick visual identification:

Opus Agents (6) -- Deep analysis & planning

AgentColorPurpose
plannerblueImplementation planning for complex features and refactoring
architectblueSystem design, scalability decisions, technical architecture
code-reviewerblueQuality, security, and maintainability review
security-reviewerredOWASP Top 10, secrets, SSRF, injection detection
tdd-guidecyanTest-driven development enforcement (RED → GREEN → IMPROVE)
database-reviewerbluePostgreSQL/Supabase query optimization, schema design

Sonnet Agents (5) -- Fast execution & automation

AgentColorPurpose
build-error-resolvercyanFix build/TypeScript errors with minimal diffs
e2e-runnercyanGenerate and run Playwright E2E tests
refactor-cleaneryellowDead code cleanup using knip, depcheck, ts-prune
doc-updateryellowDocumentation and codemap updates
verify-agentcyanFresh-context build/lint/test verification

Color Semantics

ColorMeaning
blueAnalysis & review
cyanTesting & verification
yellowMaintenance & data
redSecurity & critical
magentaCreative & research
greenBusiness & success

📋 All Claude Forge Commands

33 Commands (click to expand)

Core Workflow

CommandDescription
/planAI creates implementation plan. Waits for confirmation before coding.
/tddWrite tests first, then code. One unit of work at a time.
/code-reviewSecurity + quality check on code you just wrote.
/handoff-verifyAuto-verify build/test/lint all at once.
/commit-push-prCommit, push, create PR, optionally merge -- all in one.
/quick-commitFast commit for simple, well-tested changes.
/verify-loopAuto-retry build/lint/test up to 3x with auto-fix.
/autoOne-button automation: plan to PR without stopping.
/guideInteractive 3-minute tour for first-time users.

Exploration & Analysis

CommandDescription
/exploreNavigate and analyze codebase structure.
/build-fixIncrementally fix TypeScript and build errors.
/next-taskRecommend next task based on project state.
/suggest-automationAnalyze repetitive patterns and suggest automation.

Security

CommandDescription
/security-reviewCWE Top 25 + STRIDE threat modeling.
/stride-analysis-patternsSystematic STRIDE methodology for threat identification.
/security-complianceSOC2, ISO27001, GDPR, HIPAA compliance checks.

Testing & Evaluation

CommandDescription
/e2eGenerate and run Playwright end-to-end tests.
/test-coverageAnalyze coverage gaps and generate missing tests.
/evalEval-driven development workflow management.
/evaluating-code-modelsBenchmark code generation models (HumanEval, MBPP).
/evaluating-llms-harnessBenchmark LLMs across 60+ academic benchmarks.

Documentation & Sync

CommandDescription
/update-codemapsAnalyze codebase and update architecture docs.
/update-docsSync documentation from source-of-truth.
/sync-docsSync prompt_plan.md, spec.md, CLAUDE.md + rules.
/syncPull latest changes and sync all project docs (prompt_plan.md, spec.md, CLAUDE.md, rules). Use after any workflow or at session start.
/pullQuick git pull origin main.

Project Management

CommandDescription
/init-projectScaffold new project with standard structure.
/orchestrateAgent Teams parallel orchestration.
/checkpointSave/restore work state.
/learnRecord lessons learned + suggest automation.
/web-checklistPost-merge web testing checklist.

Refactoring & Debugging

CommandDescription
/refactor-cleanIdentify and remove dead code with test verification.
/debugging-strategiesSystematic debugging techniques and profiling.
/dependency-upgradeMajor dependency upgrades with compatibility analysis.
/extract-errorsExtract and catalog error messages.

Git Worktree

CommandDescription
/worktree-startCreate git worktree for parallel development.
/worktree-cleanupClean up worktrees after PR completion.

Utilities

CommandDescription
/summarizeSummarize URLs, podcasts, transcripts, local files.

🧩 All Claude Forge Skills

24 Skills (click to expand)
SkillDescription
build-systemAuto-detect and run project build systems.
cache-componentsNext.js Cache Components and Partial Prerendering (PPR) guidance.
cc-dev-agentClaude Code development workflow optimization (context engineering, sub-agents, TDD).
continuous-learning-v2Instinct-based learning: observe sessions via hooks, create atomic instincts with confidence scoring.
debugging-strategiesSystematic debugging techniques and profiling.
dependency-upgradeMajor dependency upgrades with compatibility analysis.
eval-harnessFormal evaluation framework for eval-driven development (EDD).
evaluating-code-modelsBenchmark code generation models (HumanEval, MBPP).
evaluating-llms-harnessBenchmark LLMs across 60+ academic benchmarks.
extract-errorsExtract and catalog error messages.
frontend-code-reviewFrontend file review (.tsx, .ts, .js) with checklist rules.
manage-skillsAnalyze session changes, detect missing verification skills, create/update skills.
prompts-chatSkill/prompt exploration, search, and improvement.
security-complianceSOC2, ISO27001, GDPR, HIPAA compliance checks.
security-pipelineCWE Top 25 + STRIDE automated security verification pipeline.
session-wrapEnd-of-session cleanup: 4 parallel subagents detect docs, patterns, learnings, follow-ups.
skill-factoryConvert reusable session patterns into Claude Code skills automatically.
strategic-compactSuggest manual context compaction at logical intervals to preserve context.
stride-analysis-patternsSystematic STRIDE methodology for threat identification.
summarizeSummarize URLs, podcasts, transcripts, local files.
team-orchestratorAgent Teams engine: team composition, task distribution, dependency management.
using-superpowersDiscover and invoke installed skills before responding to any request.
verification-engineIntegrated verification engine: fresh-context subagent verification loop.
verify-implementationRun all project verify skills and generate unified pattern verification report.

Frequently Asked Questions

What is Claude Forge?

Claude Forge is an open-source development environment for Claude Code. It bundles 11 specialized agents, 33 slash commands, 24 skill workflows, 15 automation hooks (+9 opt-in examples covering 21 lifecycle events), and 9 rule files into a single install. Think of it as "oh-my-zsh for Claude Code" -- it turns the basic Claude Code CLI into a fully equipped coding environment with built-in workflows for planning, TDD, security review, and deployment.

How is Claude Forge different from other Claude Code plugins?

Most Claude Code plugins solve one problem at a time. Claude Forge is a complete development environment -- 11 agents (frontmatter v2), 33 commands, 24 skills, 15 hooks (+9 examples), and 9 rules that work together as a cohesive system. Instead of assembling individual plugins and configuring each one, Claude Forge gives you a pre-wired pipeline: /plan feeds into /tdd, which feeds into /code-review, which feeds into /handoff-verify, which feeds into /commit-push-pr. The 6-layer security hook system also runs automatically without extra configuration.

Is Claude Forge compatible with the official Claude Code plugin system?

Yes. Claude Forge installs via symlinks to ~/.claude/ and works alongside official Claude Code plugins. Your existing settings.local.json overrides are preserved, and you can add or remove individual components without affecting the rest of the system.

How do I update Claude Forge?

Run git pull in the claude-forge directory. Because the installer uses symlinks (on macOS/Linux), updates take effect immediately -- no re-install needed. On Windows, re-run install.ps1 after pulling to copy the updated files.

Does Claude Forge work on Windows?

Yes. Run install.ps1 in PowerShell as Administrator. Windows uses file copies instead of symlinks, so you need to re-run install.ps1 after each git pull to apply updates. All agents, commands, skills, and hooks work the same on Windows, macOS, and Linux.

What does /sync do?

/sync synchronizes your project's memory and documentation. It pulls the latest changes from the remote repository and then syncs all project docs -- prompt_plan.md, spec.md, CLAUDE.md, and rule files. Run it after completing any workflow (feature, bug fix, refactor) or at the start of a new session to ensure Claude has the latest context.

How does Claude Forge handle memory across sessions?

Claude Forge uses a 4-layer memory system:

  1. Project docs (CLAUDE.md, prompt_plan.md, spec.md) -- Project-level instructions and plans that persist in the repository. /sync keeps these up to date.
  2. Rule files (rules/) -- Coding style, security, workflow conventions loaded automatically each session.
  3. MCP memory server -- A persistent knowledge graph that stores entities and relations across sessions.
  4. Agent memory (~/.claude/agent-memory/) -- Core agents record learnings after each task, improving their recommendations over time through Self-Evolution.

Running /sync at session start ensures layers 1 and 2 are current. The MCP memory server (layer 3) and agent memory (layer 4) persist automatically.


🤝 Contributing

See CONTRIBUTING.md for guidelines on adding agents, commands, skills, and hooks.



Use Claude Forge? Show it!

Add this badge to your project's README to let others know you use Claude Forge.


Contributors

claude-forge

📄 License

MIT -- use it, fork it, build on it.

If Claude Forge improved your workflow, a star helps others find it too.


Made with ❤️ by QJC (Quantum Jump Club)

// compatibility

Platformscli, api, web, mobile
Operating systems
AI compatibilityclaude
LicenseMIT
Pricingopen-source
LanguageShell

// faq

What is claude-forge?

Supercharge Claude Code with 11 AI agents, 36 commands & 15 skills — the claude-code plugin framework inspired by oh-my-zsh. 6-layer security hooks included. 5-min install.. It is open-source on GitHub.

Is claude-forge free to use?

claude-forge is open-source under the MIT license, so it is free to use.

What category does claude-forge belong to?

claude-forge is listed under automation in the Claudeers registry of Claude-compatible tools.

0 views
773 stars
unclaimed
updated 15 days ago

// embed badge

claude-forge on Claudeers
[![Claudeers](https://claudeers.com/api/badge/claude-forge.svg)](https://claudeers.com/claude-forge)

// retro hit counter

claude-forge hit counter
[![Hits](https://claudeers.com/api/counter/claude-forge.svg)](https://claudeers.com/claude-forge)

// reviews

// guestbook

0/500

// related in Developer Tools

🔓

The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Curs…

// devtoolsaffaan-m/JavaScript225,699MIT[ claude ]
🔓

Use Garry Tan's exact Claude Code setup: 23 opinionated tools that serve as CEO, Designer, Eng Manager, Release Manager, Doc Engineer, and QA

// devtoolsgarrytan/TypeScript119,234MIT[ claude ]
🔓

🙌 OpenHands: AI-Driven Development

// devtoolsOpenHands/Python79,324NOASSERTION[ claude ]
🔓

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

// devtoolsDietrichGebert/JavaScript73,042MIT[ claude ]
Connectorlinks several projects together across the ecosystem · 8 connections
→ see how claude-forge connects across the ecosystem