mindpocket

name: mindpocket description: Use when a task requires operating the MindPocket CLI to inspect server readiness, authenticate, and retrieve or manage bookmarks, folders, or the current user through structured JSON commands.

MindPocket CLI Skill

When to Use

Use this skill when the user wants to operate a MindPocket server through the mindpocket CLI instead of manual browser actions or direct API calls.

Typical requests include:

• Inspecting CLI readiness or available commands
• Configuring the target server URL
• Logging in or checking auth status
• Reading the current user profile
• Listing, reading, creating, updating, or deleting bookmarks
• Listing or reading folders

Preconditions

Check that the CLI is available before doing anything else:

command -v mindpocket

If the command is missing, tell the user the MindPocket CLI is not installed or not on PATH.

Only use a one-shot fallback such as npx mindpocket@latest ... when networked npm execution is acceptable in the current environment and the user has not prohibited it.

Never assume a default server URL. Never invent bookmark IDs or folder IDs.

Standard Workflow

Follow this sequence unless the user explicitly asks for a narrower step:

1. Confirm the CLI exists with command -v mindpocket.
2. Discover the CLI surface with mindpocket schema.
3. If a command shape is uncertain, run mindpocket <subcommand> --help.
4. Run mindpocket doctor before protected operations or troubleshooting.
5. If the server is unset, ask the user for the MindPocket server URL and set it with mindpocket config set server <url>.
6. If auth is missing or expired, use mindpocket auth login --no-open unless the environment clearly supports opening a browser.
7. Prefer read-only inspection before mutation.
8. Parse normal command output as JSON and treat failures as structured error results.

Command Selection Rules

Prefer these commands as the primary interface:

mindpocket version
mindpocket schema
mindpocket doctor
mindpocket config get
mindpocket config set server <url>
mindpocket auth login --no-open
mindpocket auth status
mindpocket auth logout
mindpocket user me
mindpocket bookmarks list --limit <n>
mindpocket bookmarks get <id>
mindpocket bookmarks create --url <url>
mindpocket bookmarks update <id> ...
mindpocket bookmarks delete <id>
mindpocket folders list
mindpocket folders get <id>

Selection rules:

• Use schema before guessing commands, arguments, or output fields.
• Use --help when a specific subcommand shape is uncertain.
• Use doctor before protected commands and before troubleshooting.
• Prefer config get before rewriting configuration.
• Prefer read-only commands before any create, update, or delete flow.

Authentication Workflow

Use this order:

1. Run mindpocket doctor.
2. If the server is missing, ask for the correct URL and set it with mindpocket config set server <url>.
3. If auth is missing or expired, run mindpocket auth login --no-open.
4. If needed, confirm status with mindpocket auth status.
5. Only use mindpocket auth logout when the user explicitly wants to clear the current session.

Prefer --no-open for terminal, remote, and agent-driven environments. Only prefer browser opening when the environment clearly supports it and the user benefits from it.

Bookmark Workflows

Use these patterns:

• To inspect recent bookmarks, start with mindpocket bookmarks list --limit <n>.
• To inspect a specific bookmark, use mindpocket bookmarks get <id>.
• To create a bookmark, require a concrete URL and use mindpocket bookmarks create --url <url>.
• To update a bookmark, first identify the bookmark ID from list or get results, then run mindpocket bookmarks update <id> ....
• To delete a bookmark, confirm intent unless the user explicitly requested deletion, then run mindpocket bookmarks delete <id>.
• To inspect folders, use mindpocket folders list and mindpocket folders get <id>.
• To inspect the current account, use mindpocket user me.

Prefer structured, read-only inspection before mutation. Do not infer IDs from names when the CLI has not returned an ID.

Failure Handling

Treat either of these as a failure:

• A non-zero process exit code
• A JSON payload with ok: false

When a command fails:

1. Surface the command that failed.
2. Surface the failure code, message, and relevant details.
3. Use mindpocket doctor if the failure might be caused by server config, auth state, or connectivity.
4. Ask for missing inputs instead of guessing them.

Common corrective paths:

• Missing CLI: tell the user to install mindpocket or approve a one-shot npx mindpocket@latest ... fallback.
• Missing server: ask for the server URL, then run mindpocket config set server <url>.
• Missing auth: run mindpocket auth login --no-open.
• Unknown arguments or flags: consult mindpocket schema or the relevant --help.

Output and Parsing Rules

MindPocket CLI is JSON-first for normal command results. Parse stdout as JSON whenever possible.

For successful commands:

• Read the JSON payload instead of scraping terminal text.
• Keep the user-facing summary concise.
• Preserve enough structured detail to debug or continue the workflow.

For failed commands:

• Include code, message, and relevant details in the response.
• Do not hide the actual reason for failure behind generic wording.

Escalation Rules

Ask the user before proceeding when:

• The server URL is unknown
• Authentication is required and the environment constraints are unclear
• A destructive action such as bookmark deletion is requested without clear intent
• The CLI is missing and using npx mindpocket@latest ... would require networked execution the user may not want

Do not silently switch to direct API calls, browser automation, or guessed parameters when the CLI flow is available but incomplete.