How to use the SeaClip-Lite Project Board CLI agent skill

What the skill lets the agent do

The skill teaches an agent to operate SeaClip-Lite through a command-line tool called cli-anything-seaclip. Once it is installed, the agent can list, create, move and delete issues; start, check, resume or stop automated pipelines; manage schedules; read the activity feed; and check server health.

The tool is stateless. It has no local state or session, communicates with the backend over HTTP, and reads SQLite directly.

• Issues: list, create, move, change status and delete
• Pipelines: start, check status, resume and stop, per issue
• Agents: list the pipeline agents
• Schedules: list, add and trigger a sync
• Activity: read recent activity
• Server: check backend health

Install the skill: download the ZIP and unzip it

Download the skill ZIP from the agent-skills directory page. The skill is a single file, SKILL.md. Unzip it into one of these locations:

For all of your projects, put the skill folder in ~/.claude/skills/. For one project only, put it in that project's .claude/skills/ folder.

You can also upload the skill in the Claude apps. Go to Settings, then Capabilities, then Skills, and upload the ZIP there.

• Personal: ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills

Install the cli-anything-seaclip package

The skill calls a command-line program, so that program must be installed too. The skill's install step is pip install -e . Run it from the package folder that contains the code.

Before you install, confirm you have Python 3.10 or newer and that the SeaClip-Lite backend is running at localhost:5200. The commands have nothing to connect to without the backend.

• Python 3.10 or newer
• SeaClip-Lite backend running at localhost:5200
• Package installed with pip install -e .

Check that the CLI can reach your board

Run a health check first. If the backend is up, the command reports its health. Then list issues to confirm you can read board data.

The skill's guidance is to add the --json flag so the output is structured for an agent to parse.

Trigger the skill with natural requests

You do not need to type commands yourself. Describe the task in plain language and the agent can choose the matching commands. Name the board items you mean, such as an issue ID or a schedule ID, so the agent does not have to guess.

Good requests say what you want and what to check. Examples of the kind of request this skill fits are below.

• Show me the 10 most recent issues with high priority and search for login bugs.
• Create an issue titled Fix export timeout with a short description and medium priority.
• Move issue ID 42 to the Review column and tell me the result.
• Check the status of the pipeline for this issue and resume it if it is paused.
• Trigger a sync for schedule 3 and show me the recent activity afterwards.

Command reference

Commands are grouped by area. Each group has its own subcommands. Flags are listed as the skill documents them.

• issue list: --status, --priority, --search, --limit
• issue create: --title, --description, --priority
• issue move ISSUE_ID --column COL
• issue status ISSUE_ID --set STATUS
• issue delete ISSUE_ID
• agent list
• pipeline start --issue UUID --mode auto or manual
• pipeline status --issue UUID
• pipeline resume --issue UUID
• pipeline stop --issue UUID
• scheduler list
• scheduler add --name, --cron, --repo
• scheduler sync SCHEDULE_ID
• activity list --limit N
• server health

Read the JSON output safely

Every command supports two output modes. The default is human-readable, with tables and colors. The --json flag produces structured JSON for agent consumption.

On failure, the skill's guidance is to check the return code, where 0 means success and anything else means an error, and to read stderr for the message. In JSON mode, errors come back as an object with an error field.

• Always use the --json flag when an agent is parsing output
• Check the return code: 0 is success, non-zero is error
• Parse stderr for error messages
• In JSON mode, errors look like {"error": "message"}

Safety notes: scripts and network

This skill runs scripts and uses the network. It connects to the SeaClip-Lite backend at localhost:5200, so the backend must be running on your machine for the commands to work.

Delete and stop actions change live project data. Deleting an issue and stopping a running pipeline are not reversible from this skill, so require a person to review those actions before the agent runs them. The skill's own description makes the same point about review.

• Runs scripts and uses the network
• Talks to localhost:5200 only as documented in the source
• Review delete and stop commands before the agent runs them
• Use a board you can afford to change while you test