cli-worker_skill

This skill delegates coding tasks to Kimi CLI agents in isolated worktrees, enabling parallel, headless task execution without disrupting the main project.
  • Python

2.5k

GitHub Stars

7

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill openclaw/skills --skill cli-worker

  • _meta.json820 B
  • AGENTS.md1.2 KB
  • CHANGELOG.md2.6 KB
  • package.json1.4 KB
  • README.md873 B
  • SECURITY.md5.2 KB
  • SKILL.md3.3 KB

Overview

This skill delegates coding tasks to Kimi CLI agents by running them in isolated git worktrees. It is designed for headless, parallel, or isolated work that should not interfere with the main repository. Requires the Kimi CLI to be installed and authenticated on the host where the skill runs.

How this skill works

The skill invokes the Kimi CLI to create an isolated git worktree and runs a single non-interactive task using the provided prompt and optional constraints. It returns the agent output, stores the worktree under a taskId, and provides commands to inspect status, list or remove worktrees, and perform cleanup. Verification only checks CLI availability; authentication must be completed interactively by the user.

When to use it

  • Delegate a single coding task to a CLI agent without manual interaction
  • Run isolated work in a separate git worktree to avoid conflicts with main repo
  • Execute parallel tasks and continue other work while agents run
  • Run headless non-interactive tasks from automation or CI
  • Avoid sessions_spawn when you need a guaranteed task execution

Best practices

  • Install and authenticate the Kimi CLI on the host before invoking the skill (user completes OAuth via /login).
  • Use cli-worker execute "<prompt>" for single tasks; include --constraint and --success for more predictable results.
  • Prefer CLI delegation over sessions_spawn for one-off executions; sessions_spawn can create idle sessions. If you spawn, verify child session activity within ~30s.
  • Use worktree list/status to inspect reports and decide whether to merge, keep, or remove a worktree.
  • Cleanup old worktrees regularly with cli-worker cleanup --older-than to avoid disk clutter.

Example use cases

  • Create a new feature file in isolation: cli-worker execute "Create hello.py" --constraint "Python 3.11"
  • Run tests and return artifacts headlessly: cli-worker execute "Run unit tests and report failures" --output-format text
  • Parallelize refactors across multiple branches by issuing multiple execute commands concurrently
  • Recover or inspect archived outputs from past tasks using cli-worker worktree list and status

FAQ

No. You must install and authenticate the Kimi CLI yourself; the skill does not store or manage credentials.

What if sessions_spawn creates a child session with zero messages?

Treat it as failed and retry using cli-worker execute. Also verify spawned sessions with sessions_list and sessions_history within ~30 seconds.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational