Learn
Skills teach your agent
how to use the capabilities it doesn't ship with.
The agent may know how to code, but it does not automatically know how to discover, install, authenticate, and invoke your capability runtime. A skill is the instruction layer that closes that gap. It tells Claude Code, Cursor, Codex, and similar agents how to connect to AnyCap without turning every setup into a manual integration job.
Keep the agent. Teach it the missing runtime.
Early Access
AnyCap is currently in early access. Capabilities shown on this page are available to early access users. Request access on GitHub to get started.
What a skill adds to the workflow
Skill
Instruction layer that teaches the agent what missing capabilities exist and how to install and invoke them.
CLI
Executable surface the agent or operator actually runs, like `anycap image generate`.
Capability
Concrete action exposed through the runtime, such as image generation, image understanding, or video analysis.
Tool
Generic execution surface. A skill can describe a tool, but it is not the same thing as the capability runtime itself.
Where skills plug in
Three ways to install the skill
Install through skills.sh
Best when the agent ecosystem already supports skills as a distribution format.
npx -y skills add anycap-ai/anycap -a claude-code -y
Install through AnyCap CLI
Useful when you want the runtime to place the skill file in a known target directory.
anycap skill install --target ~/.claude/skills/anycap-cli/
Install manually
Fallback path when you want direct control over the skill file location.
curl -fsSL https://raw.githubusercontent.com/anycap-ai/anycap/main/skills/anycap-cli/SKILL.md --create-dirs -o ~/.claude/skills/anycap-cli/SKILL.md
What skills are not
A skill is not the same as MCP, A2A, or another agent protocol. Those protocols describe how systems communicate. A skill describes how the agent should install and use a capability runtime inside its execution environment.
A skill is also not the capability itself. It is the instruction layer that points the agent toward the runtime, the CLI, and the concrete commands it can call once the install path is complete.
Where to go next
Learn
Why one CLI matters
Use this when you want to see how the command surface stays stable across image, video, and vision workflows.
Learn
What agents can't do
Use this when you want the capability-gap narrative before choosing the next product or capability page.
Guide
Install AnyCap
Use this when you want the shortest install path after you understand how the instruction layer works.