Development
Active planning documents for processkit evolution.
Introduction
processkit is a provider-neutral process and memory layer for AI-assisted projects.
It keeps durable project state in Git-reviewable files, validates that state through schemas and lifecycle rules, and exposes project workflows through Python MCP servers.
Release lines
| Line | Status | Use |
|---|---|---|
| v0.x | Stable and default | Existing projects and normal production use |
v1.0.0-alpha.5 | Exact-pin prerelease | Evaluation of the native lifecycle CLI and v1 contracts |
The v1 alpha is opt-in. It does not replace the supported v0 line, and it
must not be selected through an unverified latest URL.
v1 product boundary
| Layer | Responsibility |
|---|---|
| Native Rust CLI | Verify releases; plan, install, update, recover, verify, and uninstall project content |
| Python MCP runtime | Serve tools, validate entity operations, enforce transitions, route skills, and maintain the derived index |
| Visible content | Skills, schemas, state machines, processes, adapters, and configuration remain reviewable files |
| Project-owned state | WorkItems, Decisions, Artifacts, Notes, Logs, and local overrides belong to the consuming project |
Python is intentionally retained for MCP. The Rust CLI is the lifecycle and trust boundary, not a second MCP implementation.
Within this repository, context/ is processkit’s own dogfood project state.
src/context/ is the producer-owned payload installed into other projects.
They have different ownership and must not be merged.
Start with v1
The currently published native executable supports Linux ARM64 GNU systems. The CLI still requires an explicitly downloaded distribution directory; an online resolver and bootstrap installer are planned but not implemented.
Follow the v1 alpha installation and first-project tutorial for a verified, step-by-step installation.
Documentation map
- Getting Started covers installation and first use.
- Installer documents trust, transactions, automation, and runtime requirements.
- MCP Servers explains gateway and per-skill operation.
- Primitives describes durable project entities.
- Skills describes reusable process knowledge.
- Packages describes installable content profiles.
- v1 implementation status maps issue #135 to alpha.5 evidence and remaining gaps.
Getting Started
Install a verified processkit release and use its MCP tools.
Primitives
The project-memory entity model — the shape of a WorkItem, a DecisionRecord, an Artifact, and the rest of the durable record.
Skills
The skill package format, the category hierarchy, and the shipped catalog.
Packages
The package tiers, from a minimal bootstrap context to a fully managed workspace.
Processes
Process templates that sequence skills into a repeatable workflow.
Installer and releases
Standalone installation, trust, compatibility, and integrations.
MCP Servers
The gateway, per-skill servers, and the legacy aggregate bridge — plus what each harness supports.
Reference
apiVersion policy, ID formats, the migration guide, privacy conventions, and the v2 deliverable boundary.