How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

task - Skill - OpenSmartRoute

Name: task
Author: lobehub

Imported from lobehub/lobehub (packages/builtin-skills/src/task/SKILL.md). Install upstream with npx skills add lobehub/lobehub --skill task. Copyright stays with the author.

<task_skill_guides> You are executing a task within the LobeHub task system. Use the lh task CLI via runCommand to manage your task and related resources.

Task Lifecycle

Command	Description
`lh task view <id>`	View task details, instruction, documents, activities
`lh task edit <id>`	Update task name, instruction, status, priority
`lh task complete <id>`	Mark task as completed
`lh task comment <id> -m "..."`	Add a progress comment
`lh task tree <id>`	View subtask tree with dependencies

Working with Subtasks

Command	Description
`lh task create -i "..." --parent <id>`	Create a subtask
`lh task list --parent <id>`	List subtasks
`lh task sort <parentId> <id1> <id2> ...`	Reorder subtasks
`lh task dep add <id> <dependsOnId>`	Add dependency
`lh task dep rm <id> <dependsOnId>`	Remove dependency

Task Documents

Command	Description
`lh task doc create <id> -t "title" -b "content"`	Create and pin a document
`lh task doc pin <id> <docId>`	Pin existing document
`lh task doc unpin <id> <docId>`	Unpin document

Task Topics (Conversations)

Command	Description
`lh task topic list <id>`	List conversation topics
`lh task topic view <id> <topicId>`	View topic messages

Usage Pattern

Read the reference file for detailed command options: readReference('references/commands')
Run commands via runCommand — the lh prefix is automatically handled
Use --json flag on any command for structured output
Use lh task <subcommand> --help for full command-line help

Task Execution Guidelines

Use stable task references: Commands accept display identifiers (for example T-12) and stable ids (task_…). Prefer the id returned by a tool for subsequent calls. Use identifier short links (for example /task/T-12) in user-facing output; identifiers also remain compatible user input.
Check your task first: Use lh task view to understand the full instruction and context
Use task documents: Store outputs and deliverables as task documents
Report progress: Use lh task comment to log key milestones. To notify a member who needs to act, mention them with <mention name="Name" id="user_xxx" />, taking the id from lh workspace members (never guess it); a plain "@Name" is only text. Each mention sends a notification, so do not @ people in routine progress notes.
Respect dependencies: Check lh task tree to understand task ordering
Complete when done: Use lh task complete when all deliverables are ready
Automation tasks are the exception — NEVER complete them: a task with automation (heartbeat or schedule — shown as an Automation: line in the task context) is a recurring loop, and this run is one tick of it. NEVER run lh task complete (or set a terminal status via lh task edit --status) on that task: a terminal status permanently disarms the loop — no future tick will fire and nothing recovers it. A tick with nothing to do, or whose condition is not met yet, is still a successful run; just finish your turn and the next tick is armed automatically. Only the user retires a recurring task.
End conditions: the one exception is when the instruction itself defines when the loop ends ("until X", "stop once Y", "after Z is merged") and this tick has verified it is now met. Then run lh task complete <id> as your final step: it is finalized after this run finishes and retires the loop. Do not ask the user to stop it instead — nobody reads a tick's reply in real time. </task_skill_guides>

Install task and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

{ "ocm": "1", "id": "lobehub-lobehub-task", "kind": "skill", "name": "task", "description": "\\<task_skill_guides> You are executing a task within the LobeHub task system. Use the `lh task` CLI via `runCommand` to manage your task and related resources.", "publisher": "lobehub", "version": "1.0.0", "capabilities": { "domains": [ "general" ], "tags": [ "skill-md", "skills-sh" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "\\<task_skill_guides> You are executing a task within the LobeHub task system. Use the `lh task` CLI via `runCommand` to manage your task and related resources." ], "primary": false, "metadata": { "source": { "provider": "skills.sh", "repository": "https://github.com/lobehub/lobehub", "path": "packages/builtin-skills/src/task/SKILL.md", "ref": "HEAD", "url": "https://github.com/lobehub/lobehub/blob/HEAD/packages/builtin-skills/src/task/SKILL.md", "key": "lobehub/lobehub/packages/builtin-skills/src/task/SKILL.md" } }, "instructions": "\\<task_skill_guides>\nYou are executing a task within the LobeHub task system. Use the `lh task` CLI via `runCommand` to manage your task and related resources.\n\n# Task Lifecycle\n\n| Command | Description |\n| ------------------------------- | ----------------------------------------------------- |\n| `lh task view <id>` | View task details, instruction, documents, activities |\n| `lh task edit <id>` | Update task name, instruction, status, priority |\n| `lh task complete <id>` | Mark task as comp", "cost": { "context_tokens": 1089 } }

task

About

Task Lifecycle

Working with Subtasks

Task Documents

Task Topics (Conversations)

Usage Pattern

Task Execution Guidelines

Install task and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Use it

Manifest

Reviews