Enforced Intelligence Framework — 83 skills, 10 agents, 8 pipelines, 5 synapses, 12 bundles, runtime contracts, all with policy-gated guardrails
Install OMNISKILL, configure your platform, and start using skills and agents.
Command-line interface reference for managing skills, agents, pipelines, and bundles.
Upgrade guide from v0.2.0 — fully backward compatible, zero breaking changes.
Upgrade guide from v2.0 — runtime contracts, policy engine, telemetry. Zero breaking changes.
NEWAdapters → Skills → Bundles → Agents → Pipelines → Runtime Contracts. The full v3 blueprint.
UPDATEDReal execution engine with state machines, artifact validation, and failure recovery.
NEW10 Iron Laws, anti-rationalization, forbidden phrases, and the deviation protocol.
NEWDECOMPOSE → REASON → VALIDATE → SYNTHESIZE chain-of-thought protocol.
NEWWrite SKILL.md + manifest.yaml to add new capabilities to the framework.
Define specialized agents with guardrails, triggers, and skill dependencies.
Build multi-step workflows that chain agents with validation gates.
Package related skills into installable bundles for different workflows.
Build cognitive protocols that shape how agents think and reason.
Configure OMNISKILL for Claude Code, Copilot CLI, Cursor, Windsurf, and more.
Detailed profiles of all 10 specialized agents and their capabilities.
Full catalog of platform integrations and adapter configurations.
Frequently asked questions about OMNISKILL usage and troubleshooting.
Version history and release notes for all OMNISKILL releases.
Project overview, quick start, and feature summary.
How to contribute skills, agents, bug fixes, and documentation.