canopy-capabilities

Maps Canopy workspace capabilities to user needs and discovers the live tool surface.

Updated Jun 12, 2026
One-click install
npx skills add https://github.com/UniverLab/univerlab --skill canopy-capabilities-univerlab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: canopy-capabilities
Source: https://github.com/UniverLab/univerlab/tree/main/public/.well-known/agent-skills/canopy-capabilities
Command: npx skills add https://github.com/UniverLab/univerlab --skill canopy-capabilities-univerlab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? New users of a Canopy workspace often ask "what can I do here?" without knowing which capability — memory, background agents, loops, sync, RAG, or spec backlogs — fits their need. This Skill provides a capability map with starter ideas and teaches the agent to query the live tool surface instead of relying on a static inventory that goes stale. ## Core Features & Use Cases - Capability Map: Covers seven Canopy capabilities — persistent project memory, cron/file-watch background agents, autonomous work loops, multi-agent sync, TUI session management, RAG over local documents, and spec backlogs with pools. - Need-to-Capability Table: Translates statements like "run X every night" or "two agents keep stepping on each other" into the right Canopy feature with concrete triggers, platforms, and verification steps. - Live Surface Discovery: Instructs the agent to verify capabilities via get_tools, canopy --help, agent_models, and blueprint_list so guidance stays accurate as the tool surface evolves. - Use Case: A user asks how to stop re-explaining their codebase every session; the Skill maps this to the intelligence layer and suggests persisting facts and patterns per project. ## Quick Start Ask the agent what you can automate in this Canopy workspace and describe your recurring task or pain point.

Frequently Asked Questions about canopy-capabilities

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What can I do with a Canopy workspace?

Canopy provides persistent project memory, background agents on cron or file-watch triggers, autonomous work loops, multi-agent sync, a TUI dashboard, RAG over local documents, and spec backlogs. Use get_tools and canopy --help to see the live surface.

How do I automate a recurring task in Canopy?

Recurring tasks map to background agents, which run any prompt on a cron schedule or when files change, on any installed CLI platform. Define the trigger, the platform, a deterministic verification step, and what done looks like.

How do I run a backlog of tasks unattended with AI agents?

Use spec backlogs with pools feeding a work loop: a graph of agent, check, and gate nodes that implements, verifies with real commands, reviews on a different model, and commits. Loops survive restarts and resume after CLI quota resets.

Does this skill require the Canopy MCP server?

Yes, the skill declares requires = "canopy" and depends on the Canopy MCP server being available. Its discovery commands like get_tools, agent_models, and blueprint_list are exposed through that server.

Why does the skill not list exact tool names and flags?

Tool names and flags evolve, so a static inventory would rot. The skill teaches the agent to query the live surface first, and when the skill and the live surface disagree, the live surface wins.