Duvo Agent Skills

Official agent skills for Duvo — the AI platform that maps your real operations, designs the better version, and runs them across your existing systems.

These skills teach Claude and other coding agents to operate Duvo directly: manage Agents, debug failed Runs, and write the AOPs that drive them. How you install depends on which agent you use — see below.

There are two ways to install, depending on where you work:

• In the Claude app (plugin) — for Claude Desktop / claude.ai users. One install adds the skills and the Duvo connector; no terminal needed.
• In a coding agent (npx) — for Claude Code, Codex, Cursor, and other compatible tools. Installs skills individually or as a set, paired with the duvo CLI.

Use in the Claude app (plugin)

If you build and run your Duvo Agents from the Claude desktop app or claude.ai, install the Duvo plugin — it bundles the skills together with the Duvo connector (https://api.duvo.ai/v2/mcp), so a single install wires up both the know-how and the data access:

1. Open Customize → Plugins → "+" → Add marketplace → Add from a repository.
2. Enter duvoai/skills.
3. Install the Duvo plugin from the marketplace.
4. The first time a skill reaches for Duvo, you'll be prompted to sign in (OAuth) — no CLI, no API keys.

In Claude Code the equivalent is:

/plugin marketplace add duvoai/skills
/plugin install duvo@duvo
/reload-plugins

Notes:

• Plugin-installed skills are namespaced, e.g. duvo:run-debugger, duvo:aop-writer.
• The plugin ships every skill except duvo-cli — that one drives the terminal CLI, which doesn't apply inside the Claude app. Use the npx path below if you want it.
• Already added the Duvo connector manually? The plugin brings its own plugin-scoped copy of the same tools — remove the manually-added connector after installing, so you don't end up with two copies of every Duvo tool.
• Updates: when a new plugin version is released, the plugin manager shows an Update button — one click and you're current.

Quickstart (coding agents)

Get from zero to a working setup in four steps — install the skills, add the CLI they call, and sign in:

# 1. Install the skills into your agent
npx skills add duvoai/skills

# 2. Install the Duvo CLI (used by the skills to talk to Duvo)
npm install -g @duvoai/cli

# 3. Authenticate — opens your browser for OAuth sign-in
duvo login

#    ...or sign in with an API key from https://app.duvo.ai/settings/api-keys
#    duvo login --api-key dv_live_...

# 4. Verify you're signed in
duvo whoami

Then just ask your agent — for example: "Why did Run run_abc123 fail?", "Rewrite this AOP to be clearer", or "List my failed runs from yesterday".

Install a specific skill

npx skills add duvoai/skills --skill duvo-cli

Install for a specific agent

Use --agent to target one or more agents (you can pass several, space-separated):

# Claude Code
npx skills add duvoai/skills --agent claude-code

# Codex
npx skills add duvoai/skills --agent codex

# Cursor
npx skills add duvoai/skills --agent cursor

# Multiple agents at once
npx skills add duvoai/skills --agent claude-code codex cursor

Update installed skills

Skills installed via npx don't auto-update — run npx skills update to pull the latest versions.

Available skills

Each skill is self-contained — install the whole set or just the one you need. All of them are available via npx; the Claude-app plugin ships all except duvo-cli.

Skill	In plugin	Description
duvo-cli	—	Drive the Duvo platform from the terminal via the duvo CLI (@duvoai/cli) — manage agents, runs, cases, queues, files, skills, connections, and Clarity processes, or hit any endpoint with duvo api.
aop-writer	✓	Draft, rewrite, or critique a Duvo Agent AOP, returning a single Markdown document in the canonical GOAL / STEPS / NOTES shape.
run-debugger	✓	Investigate why a Duvo Run failed or produced the wrong outcome — reads the Run transcript and active Build, names the root cause from a fixed failure-mode taxonomy, and proposes one concrete fix.
workflow-debugger	✓	Audit a Duvo Agent — or a multi-Agent workflow connected by a Queue — across many Runs to find systemic inefficiencies and quality issues, then recommend concrete AOP and architecture changes.
connection-doctor	✓	Diagnose and health-check the Connections and credentials a Duvo Agent relies on — reads each Build's Connections and their live state, names the problem, and proposes one concrete fix.
improve-agent	✓	Run the guided, end-to-end loop that makes one Duvo Agent better — survey its setup, ground findings in recent Runs via the debugger skills, propose concrete changes, and apply them on your yes.
improve-queue	✓	Run the guided, end-to-end loop that makes one Duvo Queue — and the producer→consumer workflow around it — better, ending in applied AOP, topology, and trigger changes.

Learn more

• Duvo website
• Documentation
• Duvo MCP server — the connector the plugin bundles
• Duvo CLI on npm

License

MIT