Enterprise Skills CLI

Harness engineering for AI coding agents: 93 governed skills, deterministic safety gates, audited multi-skill workflows, and compliance evidence — across Cursor, VS Code, Claude Code, Windsurf, JetBrains, Neovim, Xcode, and plain CLI.

Your IDE's AI does the work. The Enterprise Skills engine decides what runs, blocks what shouldn't, and records tamper-evident proof of everything.

Install

npm install -g enterprise-skills

Quickstart

# Have a license key? (free-tier keys available at https://enterpriseskills.ai)
enterprise-skills license activate <your-key>   # downloads your skill catalog

# Set up a project (renders IDE adapters, authority pack, orchestration hooks)
enterprise-skills onboard --with-init --install-cursor-hooks

# Not sure what to run? Ask the engine.
enterprise-skills orchestrate next

The agentic layer

# Deterministic skill resolution for any task — same input, same answer
enterprise-skills preflight "fix the auth timeout before we ship"

# Multi-skill pipelines with evidence gates (12 built-in workflows)
enterprise-skills orchestrate workflow start release-readiness
enterprise-skills orchestrate workflow next
enterprise-skills orchestrate workflow complete <step-id>

# Fully autonomous, gates as hard stops (explicit opt-in; bring your own agent)
enterprise-skills orchestrate workflow run release-readiness --yes \
  --agent-cmd "claude -p \"{prompt}\""

# Governance control detection with per-control evidence (all tiers)
enterprise-skills compliance controls

# SOC 2 / HIPAA / GDPR / ISO 27001 / PCI DSS / FedRAMP mapping (Enterprise tier)
enterprise-skills compliance report

# Deterministic Release Governor over the actual diff — classify base...HEAD
# risk, bind skill-output evidence to the commit, PASS/FAIL (exit 1 on FAIL);
# --post completes the PR's required check run
enterprise-skills govern --post --pr 42

The Governor consumes evidence from .project-ai/skill-outputs/ — run the domain audit skills (or orchestrate workflow run release-readiness) for the current commit first. Missing evidence for a required domain fails closed by policy; evidence produced for an earlier commit is rejected as stale.

Every gate decision, workflow transition, and report lands in an append-only SHA-256 audit chain under .project-ai/ — execution evidence your reviews and audits can cite.

Customer Evidence Center (Team and Enterprise)

Cloud synchronization is always opt-in per project. Preview the exact allowlisted text artifacts before anything leaves the machine:

enterprise-skills evidence status
enterprise-skills evidence preview
enterprise-skills evidence sync
enterprise-skills evidence open
enterprise-skills evidence revoke <share-id>

The collector includes orchestration audit events, generated skill outputs, and LaunchOps audit reports. It excludes source code, .env files, credentials, and arbitrary attachments. Team bundles retain for 90 days; Enterprise bundles retain for 365 days unless a custom policy applies.

Tiers

| Tier | Skills | Highlights | |------|:------:|-----------| | Free | 8 | Evaluation: core lifecycle + review skills, preflight resolver | | Pro | 90 | Full catalog, all audits, updates | | Team | 90+ | Per-seat licensing, governance features, all IDE adapters | | Enterprise | all | Compliance mapping, control plane, sales-led |

Pricing and docs: https://enterpriseskills.ai · Getting started: https://enterpriseskills.ai/getting-started

Telemetry

Anonymous usage telemetry can be disabled with ENTERPRISE_SKILLS_TELEMETRY=off. License validation traffic is required for paid content delivery. Privacy policy: https://enterpriseskills.ai/privacy

© ArcFusion Technologies Corp. License: see LICENSING terms shipped with the package and https://enterpriseskills.ai/terms