Runtime Skill Synthesis Protocol
Transform your agents from completion-oriented tools into growth-oriented experts. IKKF distills your experience into your own sovereign, reusable skills — built from your work, for your setup, under your control. No third-party skill blobs you can't audit.
Solving the problem is only half the battle. Remembering how to solve it is the victory.
Most AI agents are optimized for the "Final Result." They execute a task, deliver the output, and then discard the process. This creates Knowledge Debt — the loss of expertise the moment a session is compacted or restarted.
IKKF introduces a hard systemic block: The Synthesis Gate. A task cannot be marked as complete until the agent distills the experience into a reusable SKILL.md. We transform "doing the work" into "building an expert system."
Every skill is yours — generated from your own tasks, for your own workflow and setup. You can read, audit, and control every line. No opaque, externally-created skill that might be doing something you never approved. Your expertise stays yours.
Point IKKF at your AI agent, then watch the seven phases execute end-to-end
Each belt represents a level of mastery in the IKKF protocol
Every skill is generated from your own tasks, for your workflow and setup — plain-text SKILL.md you can read and audit line by line. No opaque, externally-created skills that might be doing something you never approved.
IKKF pairs with your AI agent to run every phase. Point it at OpenAI, Claude, Ollama, Gemma, or any OpenAI-compatible gateway.
One protocol, every agent. Claude, Hermes, OpenClaw, local models, or your own gateway — swap providers in a minute.
Runs on your machine with no cloud dependency and no telemetry. Use local models for full privacy, or cloud agents for speed.
Not just tracking — the full journey runs end-to-end: Acknowledge → Detect → Check → Research → Decompose → Synthesize → Execute.
Light mode for instant scaffolding, or Full mode where the agent does deep research and builds real deliverables.
Phase 7 writes actual files — apps, docs, and reusable skills — into your workspace, not just plans on a screen.
Hard-block on task completion. Expertise is distilled into a permanent skill before a task closes. No knowledge debt.
Works vanilla in any setup. Point IKKF at any host, port, or model — nothing is hard-wired.
IKKF transforms uncertainty into structured execution
State IKKF activation. Set expectations with the user.
Identify core technology from task description in under a minute.
Search for existing skills. Verify currency of knowledge.
Web search for 2026 best practices. Load current expertise.
Break into 10-15 minute chunks. Estimate time for each.
Create SKILL.md with framework, code samples, and pitfalls.
Complete chunks, update status, archive to long-term memory.
No license. No paywall. The full IKKF protocol, yours to keep.
The entire v3 protocol, unlocked forever. No machine-lock, no expiry, no paywall.
IKKF v3 is free forever — this isn't a trial or a teaser. If it saves you time, buy me a coffee — or fund the next build.
Your support keeps IKKF free and open for everyone. Every coffee goes back into the build.
A native desktop app that brings the whole protocol to your fingertips — no terminal, no configuration dance.
A first-class macOS app — built in Swift, not a wrapper — with more platforms on the way.
Point it at OpenClaw, Claude, OpenAI, or local Ollama and it just works. No base URLs, no hunting for keys.
Every skill you synthesize, browsable and auditable in-app — built from your work alone.
Your data stays on your machine. No cloud, no telemetry.
On pricing: v3 stays free forever, as promised. v4 is being built to be worth paying for — however it ships, the people who support IKKF now help decide the shape of it.
One command to install. No dependencies required.
# One-line install — free, no license required curl -fsSL https://ikkf.info/install.sh | bash # Or download and verify curl -fsSL https://ikkf.info/install.sh -o install.sh bash install.sh # Initialize and start ikkf init ikkf start
# In PowerShell (Admin) irm https://ikkf.info/install.ps1 | iex # Or download and run # Download install.ps1, then right-click → "Run with PowerShell" # Initialize and start ikkf init ikkf start
# Download the CLI package curl -fsSL https://ikkf.info/ikkf-cli.zip -o ikkf-cli.zip unzip ikkf-cli.zip -d ikkf-cli cd ikkf-cli # Create virtual environment python3 -m venv venv source venv/bin/activate # Linux/Mac # or: venv\Scripts\activate # Windows # Install dependencies pip install -r requirements.txt # Run python3 -m ikkf --help