v5.15.0 • Production Ready

The Forge
for Production AI Development

Standalone CLI or IDE skill layer — your choice. 127+ skills, 22 real tool agents, Semgrep SAST, PRD linting, cross-platform parity detection, and a self-improving knowledge base. Quality gates your AI can't skip, built in from day one.

64
CLI Skills
22
Tool Agents
127+
Total Skills
6
AI Providers

Standalone + Five IDEs

Run sf on its own with any API key, or install skills into the IDE you already use. Same agents, same gates, everywhere.

Standalone CLI

npm install -g skillfoundry && sf setup — full pipeline with no IDE. Interactive setup wizard, 6 providers, budget controls, runs in any terminal.

🤖

Claude Code

62 command skills via /command syntax. Original platform with full feature support including autonomous mode.

🐙

GitHub Copilot

64 custom agents with GitHub integration. PR reviews, CI/CD debugging, and more.

Cursor

60 rules for AI assistance. Automatic loading and intelligent routing.

🔮

OpenAI Codex

60 skills via $skill-name invocation. Native Skills format.

💎

Google Gemini

61 generated skill documents with platform-specific optimizations.

Forged for Production

Every feature is battle-tested and production-ready. No compromises.

Standalone or IDE — Your Choice

sf setup gets you running in under a minute with no IDE. Or install skills into Claude Code, Cursor, Copilot, Codex, or Gemini. Same pipeline, same gates, either way.

No Lock-In
🛡

The Anvil Quality Gate

8-tier validation (T0-T7) between every agent handoff. Correctness, security, types, tests, build, scope, deploy pre-flight. Gates run in the CLI too — not just in IDEs.

Zero Defects
😈

Specter Security Engine

Adversarial threat modeling runs automatically in every pipeline. Generates attack vectors with CVSS metadata and simulates exploits before your code ships.

Red Team Built-In
📝

PRD-First Pipeline

Drop requirements in genesis/, run /go or sf forge — validates DB, Backend, Frontend automatically. No vibe-coding, no drift from spec.

Genesis-First
📊

Semantic Vector Memory

Every pipeline run is semantically indexed. Hybrid recall combines keyword + cosine similarity search. Local-first with Transformers fallback — no API key needed for memory.

AI Memory
🎯

Autonomous Developer Loop

Auto-classifies every input: FEATURE/BUG/REFACTOR/QUESTION. Routes to correct pipeline without manual command invocation.

Zero Friction
🔒

6 Providers, Budget Controls

Anthropic, OpenAI, xAI, Google, Ollama, LM Studio. Per-run and monthly cost caps. Auto-selects the first available provider at startup.

Provider Agnostic
💻

VS Code Extension

Setup wizard stores your API key in SecretStorage. CLI check offers npm install if sf is missing. Forge progress notification stays open while the pipeline runs. Gates, dashboard, CodeLens — no terminal required.

Native IDE Shell

See It In Action

From idea to production in seconds. Watch the framework in action.

$npm install -g skillfoundry
$cd ~/DevLab/MyProject && sf setup
No API key detected. Choose a provider:
1) Anthropic 2) OpenAI 3) xAI 4) Google 5) Ollama
Select provider [1]: 1
Paste your API key: sk-ant-...
✔ Anthropic configured. Ready.
$sf forge
⚙️ Validating PRDs in genesis/...
🔨 Generating stories...
⚡ Pipeline: IGNITE → PLAN → FORGE → INSPECT → SPECTER → DEBRIEF
😈 Specter: 0 critical vectors found
🛡 Anvil: T0-T7 validation passed (8/8)
✅ Backend tests: 47/47 passed
✔ Production-ready code generated

Ready to Forge Production Code?

Join developers using SkillFoundry to build production-ready applications with AI assistance.