Fusion Skills
Route one request into Fusion’s skill lifecycle—discover, install, sync, greenkeep, author, or triage—via MCP or catalog fallback.
Overview
fusion-skills is an agent skill most often used in Build (also Operate iterate) that routes Fusion skill lifecycle intents to the correct specialist agent via MCP or catalog fallback.
Install
npx skills add https://github.com/equinor/fusion-skills --skill fusion-skillsWhat is this skill?
- Single entrypoint intent router for find, install, update, sync, greenkeep, setup automation, create skill, and report s
- Loads only the routed agent file (discovery, greenkeeper, author, warden)—references such as skill-catalog.md on MCP fai
- Primary integration: mcp_fusion_skills with promoted fallback to references/skill-catalog.md when Fusion MCP is unavaila
- Explicit DO NOT USE boundary: GitHub issue resolution, PR review, task breakdown, and issue authoring belong to other Fu
- MIT-licensed orchestrator (metadata version 0.0.0) owned by @equinor/fusion-core
- Routes to 4 named agent files: discovery, greenkeeper, author, warden
- Metadata version 0.0.0, MIT license
- Suggested MCP tool: mcp_fusion_skills
Adoption & trust: 531 installs on skills.sh; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You maintain dozens of agent skills and waste time opening the wrong Fusion doc when you only meant to install, sync, or author one capability.
Who is it for?
Teams and power-user solo builders already on Fusion who need centralized skill install, sync, greenkeeping, and authoring without manually picking agent files.
Skip if: Resolving GitHub issues, reviewing pull requests, planning task breakdowns, or authoring GitHub issues—those flows are explicitly out of scope for this entrypoint.
When should I use this skill?
Finding, installing, updating, syncing, or greenkeeping skills; setting up skill automation; creating or authoring a new skill; or reporting a bug with a skill—within Fusion, not for GitHub issue/PR work.
What do I get? / Deliverables
One routed Fusion agent runs the right lifecycle operation—discovery, greenkeep, author, or warden—with MCP-first tooling and on-demand references only when needed.
- Routed execution via the correct Fusion agent (discovery, greenkeeper, author, or warden)
- Installed, synced, or authored skill artifacts per the engaged agent’s workflow
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Builders curate and maintain agent skills during product development; Fusion-skills is the orchestration entrypoint for that toolchain. Agent-tooling is where skill discovery, installation, sync automation, and authoring agents attach to your coding agent setup.
Where it fits
Discover and install a missing Fusion skill into your repo before starting a feature that depends on it.
Configure skill sync automation so teammate agents pull the same SKILL.md versions from Fusion MCP.
Run greenkeeper-style updates when upstream skills change so production agent prompts do not drift silently.
Start authoring a new internal skill and route to the author agent instead of copying template files by hand.
How it compares
Fusion lifecycle orchestrator skill—not a replacement for GitHub-centric PM skills or raw skills.sh browsing without Fusion MCP.
Common Questions / FAQ
Who is fusion-skills for?
Developers using Equinor Fusion who treat agent skills as managed dependencies and want a single SKILL.md entrypoint before install, sync, or authoring work.
When should I use fusion-skills?
During Build when setting up agent-tooling, at Operate when greenkeeping or syncing skill versions, or whenever you need Fusion discovery/install/author flows—not for GitHub issue or PR workflows.
Is fusion-skills safe to install?
It drives MCP and filesystem-backed catalog references; review Prism’s Security Audits panel and your Fusion MCP permissions before allowing install or sync operations on production repos.
SKILL.md
READMESKILL.md - Fusion Skills
# Fusion Skills Skill lifecycle entrypoint. Resolve intent, engage the right agent. ## Loading behavior Load ONLY the routed agent file. References are loaded on-demand by the agent when it needs them (e.g., `skill-catalog.md` only on MCP failure, `sync-workflows.md` only in setup mode). Do not preload all agents or all references. ## MCP tools This skill uses `mcp_fusion_skills` — the Fusion MCP tool that handles both skill discovery and lifecycle operations: - **Discovery** (`intent: query`): semantic search over the skills index by description or task. - **Inventory** (`intent: inventory`): list installed skills. - **Lifecycle** (`intent: install | update | remove`): advisory commands for managing skills. ## Routing | Intent | Agent | |--------|-------| | Find or discover the right skill; list installed skills | `agents/discovery.agent.md` | | Install, update, remove, or check skills; set up automation | `agents/greenkeeper.agent.md` | | Create, author, or improve a skill | `agents/author.agent.md` | | Inspect a skill for quality issues, or report a skill failure | `agents/warden.agent.md` | If the user was routed here from the `fusion` main gate with a partially-resolved intent (e.g., "skills"), proceed directly to the most likely agent without re-asking the top-level question. If intent is genuinely unclear and the user was not already asked a clarifying question, ask one question: "Are you looking to find a skill, install or update one, create your own, or report a problem with one?" **Compound intents:** If the request spans multiple agents (e.g., "inspect this skill and fix the issues"), run agents sequentially. Pass the output of the first agent as context to the second. **Loop prevention:** If you were activated from the `fusion` main gate, do not redirect back to `fusion`. Handle the request within your agents or state that it is out of scope. **Out of scope:** Resolving GitHub issues, reviewing PRs, planning task breakdowns, or authoring GitHub issues. Direct the user to the relevant installed skill or the `fusion` main gate — but never re-route back to `fusion` if you were activated from there. ## Migration from deprecated skills If `fusion-discover-skills` or `fusion-skill-self-report-bug` is installed alongside this skill, tell the user: > "You have a deprecated skill installed that's now part of `fusion-skills`. You can safely remove it: `npx -y skills remove <deprecated-skill-name>`" ## Safety - No secrets or credentials. - No GitHub mutations without confirmation. - No remote script execution. - No invented skill names or catalog results. # Follow-up Questions One question per agent, used only when intent is genuinely ambiguous. Ask nothing else until the answer resolves the route. Do not double-question the user — if the `fusion` main gate already asked a question, skip the top-level question here. ## Top-level routing (agent unclear) Only if the user was NOT already routed f