AIWG - AI Works Good

For each package, we recorded usage context, lockfile version, integrity, npm registry signature evidence, lifecycle scripts, native/binary surfaces, dependency-source findings, and follow-up routing. Detailed report github.com/jmagly/aiwg/tr…

External npm supply-chain audit pass for AIWG. 35 packages across runtime deps, optional/native deps, dev tooling, and shipped nested manifests.

AIWG's docSync tools help you keep your specs up-to-date with your code. Multiple modes, including 2-way sync.

AIWG power move

iykyk

New beginner docs are up for provider handoff, recovery, and onboarding validation. Get the patch: npm install -g aiwg@2026.5.10 Step into your workspace: 🔗 aiwg.io

Say hello to a smoother first run! 🚀 AIWG v2026.5.10 is live with a guided Onboarding Wizard & refreshed docs. 👇

🔄 aiwg use all now correctly restores its full behavior, deploying every framework, addon, and extension by default. 🤖 Automation is safer: wizard plans stay plan-only in JSON outputs, and script executions now require an explicit --non-interactive flag.

🧙‍♂️ aiwg wizard now guides you step-by-step through provider choice, profile selection, execution modes, and workspace verification. 🚦 aiwg status --probe --json provides clear, deterministic status reports with actionable repair guidance if configurations are malformed.

Bring secure, autonomous coding agents to your edge hardware today. Get the binaries, check out the SBOM, or spin up the versioned containers here: 🔗 github.com/jmagly/agentic… #OHHCrew #LocalAI #AIWG

🦀 Pure-Rust TLS stack (Rustls) for seamless cross-compilation 📦 Explicit versioned container tags (no more pinning to :latest) 🔒Hardened supply chain security (Cosign, SBOMs, GPG detached signatures) 🛠️ Zero-friction maintainer version-bumping tooling

The wait is over. 🚀 agentic-sandbox v2026.5.3 "The Artifact Release" is officially live! Ready for production. Built for a true, local-first AGENTIC ECOSYSTEM—no background daemons, no telemetry, just secure, isolated KVM microVMs running on your hardware. 👇

If the model is making poor decisions, the context did not contain what was required to close the gaps so it riffed. I like to frontload designs with HITL then let the agent run after having made those key decisions. This signifcantly reduces thrarsh, error and token use.

only 11 sessions, slacking

aiwg stork.ai/en/aiwg

Getting started: npm i -g aiwg cd your-project aiwg use sdlc --provider cursor # or copilot, openai, factory Docs: github.com/jmagly/ai-writ… ⭐ if useful!

🚀 AIWG 2024.12.4 "Universal Providers" is here! One toolkit. Six platforms. Deploy AI-powered SDLC workflows to: → Claude Code → GitHub Copilot → Cursor → OpenAI Codex → Factory AI → OpenCode Plus: New testing-quality addon with TDD npm i -g aiwg

The testing-quality addon includes: - tdd-enforce — pre-commit hooks + CI coverage gates - mutation-test — validate tests beyond coverage (Stryker/PITest) - flaky-detect — find unreliable tests from CI history - flaky-fix — pattern-based auto-repair