← Back to Help

Team Governance Playbook

Rolling out governed AI coding across a team with Enterprise Skills.

1Why governed AI

Native IDE AI is a powerful engine, but it starts from zero every session, drifts across models, and offers no objective signal that code is safe to ship. Governance turns that raw capability into something consistent, reviewable, and accountable — without slowing developers down.

2Roll it out

Install the CLI wherever developers work: npm install -g enterprise-skills.

Initialize per repo with enterprise-skills init — it auto-detects each developer’s editor and renders the same governed skills into all of them, so standards don’t depend on tool choice.

Pick a governance profile that matches the repo’s risk (e.g. standard vs. regulated). The profile constrains how agents work in that codebase.

3Standards & the authority pack

Every repo gets a .project-ai/ authority pack — your patterns, standards, and open threads. It loads automatically each session so agents build from your source of truth instead of guessing. Review .project-ai/PROJECT_PROFILE.yaml after init and keep it honest; it’s the highest-leverage file in the repo.

4Workflows & release gating

Use agentic workflows for multi-step, gated pipelines instead of one-shot chat:

enterprise-skills orchestrate workflow start release-confidence — deterministic scoring returns go / conditional / no-go before you ship, with the reasons attached.

enterprise-skills preflight "…" resolves a task into a plan; compliance report produces an audit-ready summary.

5Teams & seats

On Team and Enterprise tiers, everyone inherits the same governed standard — no per-developer drift. Seats are managed centrally; new hires get the full governed setup from their first init, not a blank slate.

6Compliance & audit

Governed sessions produce an audit trail and SOC 2-ready reports — evidence you can hand to security and procurement. Native AI produces nothing here; this is often the difference between “allowed in the SDLC” and “not.”

7Adoption tips

Start with one or two repos and a light profile; expand as trust builds. Coach the team to describe outcomes, not keystrokes, and to let skills auto-route. Make session start and a release workflow the two non-negotiable habits — everything else follows.

You keep your model and your editor. Enterprise Skills makes them consistent, safe, and accountable — in every IDE, on every model.

enterpriseskills.ai · Getting Started