AI-DLC Engineering OS 3.1.0 — one-command setup

One-command setup

With uv and Git installed, run:

uvx --from git+https://github.com/jcamargos/ai-dlc-engineering-os@v3.1.0 ai-dlc-engineering-os setup

Choose Codex, Claude Code or VS Code, review the plan and confirm. Setup installs a persistent uv tool environment, backs up configuration and memory, registers MCP and verifies the selected editors. Restart the editors afterward.

• Preview changes with --editors codex,claude --dry-run.
• Unattended installation requires explicit --editors and --yes.
• No read-only tool permissions are pre-approved by the new setup.
• The MCP runtime survives clearing the uvx cache.
• Existing source installers remain available; installation/update/uninstall details are in docs/INSTALLATION.md.

The short uvx ai-dlc-engineering-os setup command is not available yet: PyPI account and Trusted Publisher setup are pending. A manual OIDC publication workflow and exact configuration instructions are included.

Validation: 82 Python tests and 14 subtests passed locally; Ruff and both validators passed. CI checks the built wheel through uvx and then verifies MCP after cache cleanup, using isolated homes. Required PR CI passed before merge.

MIT license. No changes to earlier release tags.

AI-DLC Engineering OS 3.0.2 — MIT license and refreshed README

What's changed

• MIT License: project licensing is aligned across Python, SDK and plugin metadata. Third-party dependencies retain their own licenses.
• Refreshed README: centered navigation, an original workflow diagram, collapsible FAQ, documentation index and Portuguese setup guidance.
• Removed agentshield.com.br references from the README and NOTICE, and clarified local memory storage versus the AI client's model-provider data handling.
• Aligned package, CLI, SDK, plugin and marketplace version identifiers to 3.0.2.

Earlier release tags retain their original license notices. This release does not change the default runtime behavior.

Validation

74 Python tests and 14 subtests, Ruff, agent registry and platform validators, TypeScript build and 2 SDK tests passed locally. Required pull-request CI passed before merge.

Install

Download the source archive below, or clone this release:

git clone --branch v3.0.2 https://github.com/jcamargos/ai-dlc-engineering-os.git
cd ai-dlc-engineering-os

Follow the README's Windows or macOS/Linux installation instructions.

AI-DLC Engineering OS 3.0.1 — free and local by default

100% free and local by default.

• No implicit remote endpoint. The Python and TypeScript SDK clients and agentshield init no longer default to any hosted server. Pass an endpoint explicitly or set AI_DLC_REMOTE_URL; without one, init configures a local-only project. Breaking only for callers that relied on the implicit default.
• Free and open source: no hosted or commercial offering referenced; contributions now use the DCO sign-off (git commit -s) instead of a CLA.
• New README: requirements, what the installer changes in each host, the 26 agents, commands, security model, FAQ and uninstall.
• Test fixtures no longer reference private repository names.

Upgrade: git pull in your clone and re-run the installer (./install/install.sh or install\install-windows.ps1).

AI-DLC Engineering OS 3.0.0 — first open-source release (Apache-2.0)

Initial public release of AI-DLC Engineering OS, the local edition of AgentShield.

• Governed engineering memory (SQLite/WAL): FTS5/BM25 and local vector search, decisions,
  lessons, incidents, outcomes, versioned in-place migrations, hard purge, project merge.
• 26 specialist agents with contracts, evaluation gates, staged releases, receipts and drift.
• Human-only governance through ai-dlc-admin; credentials refused on every write.
• Persistent swarms with leases and worktrees; Mission Control; opt-in federation/marketplace.
• ai-dlc-install: one-command global setup for Codex, Claude Code (MCP, CLAUDE.md, 26 subagents)
  and VS Code (GitHub Copilot agent mode); doctor and clean uninstall.
• Native plugin for Codex and Claude Code; Python and TypeScript SDK clients.

Install

git clone https://github.com/jcamargos/ai-dlc-engineering-os && cd ai-dlc-engineering-os
./install/install.sh   # macOS / Linux
powershell -ExecutionPolicy Bypass -File .\install\install-windows.ps1   # Windows
ai-dlc-install doctor

License: Apache-2.0. Team edition: AI-DLC OS Cloud by AgentShield.