Core Platform
Skills and process management for the AI era
A platform where every skill is governed as an operational process with a human-readable form, a named owner, full version history, and an MCP interface. Publish once, then distribute the right skills to the right people and AI agents.
Platform capabilities
Everything your skills and processes need to stay alive
Markdown + YAML frontmatter
Every skill is governed as a human-readable Markdown process document with structured YAML frontmatter, following the Agent Skills spec. Editable by anyone, parseable by any AI.
Immutable version history
Every save creates a new immutable version. View diffs, roll back to any point, and see which version an AI client consumed on any given call.
Named ownership
Every department, skill, and process has a named owner, a real person accountable for keeping it accurate. Ownership is the accountability anchor.
MCP delivery and agent distribution
An MCP server delivers skills to compatible AI clients. Admins control which automated workers can use each skill without redeploying agent code.
Usage analytics
Every skill use writes a structured event. See adoption rates, per-client and per-agent breakdowns, staleness pressure, and health scores across skills and processes.
Staleness alerts
Configurable rules trigger email or webhook notifications when a skill or process hasn't been reviewed in N days, or has been called many times without an update.
Org hierarchy
Structure that mirrors how your org works
Koinoflow organizes skills and processes around a four-level hierarchy that maps naturally to how real companies operate. Ownership cascades down: every department and every governed process has a named accountable person.
Workspace
Acme Corp · Top-level org unit
Team
Engineering · Division or business unit
Department
Frontend · Functional area with an owner
Process
Deploy to production · Owned procedure, the atomic unit
Usage event
Skill: Deploy to production
Version: 2.4.1
Used by: Release agent
Time: 2026-04-11 09:12
30-day analytics
Skill: Deploy to production
Total uses: 847
Agent uses: 219
Last reviewed: 86 days ago
Health score: 34
Alert: Review overdueObservability
Turn AI usage into actionable insight
Every skill use writes a structured usage event. The analytics dashboard turns that raw log into adoption rates, staleness pressure, per-client and per-agent breakdowns, and health scores across your skills and processes.
- "Your onboarding skill was called 847 times last month"
- "The release agent read the deploy skill 219 times. No other agent can see it."
- "It hasn't been updated in 90 days. Here are the 3 AI clients still on v1.2."
- "Your Deploy skill has a health score of 34. It's stale and unowned."
Included in all plans.
Give your AI agents a governed system for skills and processes
Start with a free workspace. MCP server running in 30 minutes.
Open source (MIT) · free to self-host · managed hosting by Visionect