Skip to content
Agentmart
Submit a tool

orchestration-run-task

Use when Codex or another manager needs to hand off a concrete subtask to opencode, ollama, or gemini through scripts/orchestration/run_worker.sh.

Skill by haitaopanhq

IndexedApache-2.0★ 1
Write a review

Chalk Labs

Need help deploying orchestration-run-task?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/haitaopanhq-orchestration-run-task

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when Codex or another manager needs to hand off a concrete subtask to opencode, ollama, or gemini through scripts/orchestration/run_worker.sh. Defines the worker execution contract, prompt shape, handoff rules, and fallback behavior.

Use cases: documents, productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/orchestration-run-task

What's inside

Skill name
orchestration-run-task
Description
Use when Codex or another manager needs to hand off a concrete subtask to opencode, ollama, or gemini through scripts/orchestration/run_worker.sh. Defines the worker execution contract, prompt shape, handoff rules, and fallback behavior.
License
Apache-2.0
Instructions
211 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • body: contains an absolute path from the author's machine

Files (2)

  • SKILL.md
  • manager-integration.md · references/

SKILL.md (first lines)

# Orchestration Run Task

## Goal

Use [scripts/orchestration/run_worker.sh](/Users/shenlan/workspaces/cloud-neutral-toolkit/github-org-cloud-neutral-toolkit/scripts/orchestration/run_worker.sh) as the single worker execution entrypoint.
Use [config/orchestration/roles.yaml](/Users/shenlan/workspaces/cloud-neutral-toolkit/github-org-cloud-neutral-toolkit/config/orchestration/roles.yaml) as the single source of truth for manager and worker roles.
Use [scripts/orchestration/run_manager.sh](/Users/shenlan/workspaces/cloud-neutral-toolkit/github-org-cloud-neutral-toolkit/scripts/orchestration/run_manager.sh) when a manager should load skills plus role metadata without hand-assembling prompts.
Use `--prompt-mode compact` as the default manager mode. Use `--prompt-mode full` only when debugging or when the compact contract is insufficient.
Use `--template review|deploy|dns-audit` to bind compact mode to a fixed output contract.

This skill is the execution layer under [../codex-multi-cli-orchestrator/SKILL.md](../codex-multi-cli-orchestrator/SKILL.md):

- `codex-multi-cli-orchestrator` decides what to delegate
- `orchestration-run-task` turns that decision into a concrete worker command

Use this skill when:

- a manager must dispatch one bounded subtask to one worker
- the worker prompt must stay short and structured
- the worker result must come back to Codex or another manager for acceptance

## Preconditions

Before using this skill, verify:

1. the worker CLI exists locally
2. the needed model credentials are available
3. the task can be bounded to one worker
4. the target role can be resolved from `config/orchestration/roles.yaml`

Read [references/manager-integration.md](references/manager-integration.md) only when you need manager-side integration patterns for `opencode`, `Antigravity`, or another orchestrator.
Use [scripts/orchestration/orchestration_roles.py](/Users/shenlan/workspaces/cloud-neutral-toolkit/github-org-cloud-neutral-toolkit/scripts/orchestration/orchestration_roles.py) to inspect the configured role map.

## Entrypoint

Use exactly one of these commands:

```bash
scripts/orchestration/run_worker.sh --describe opencode
scripts/orchestration/run_worker.sh opencode "<prompt>"
scripts/orchestration/run_worker.sh ollama "<prompt>"
scripts/orchestration/run_worker.sh gemini "<prompt>"
```

Use `codex` only when the orchestrator explicitly falls back to Codex direct execution.

The wrapper already reads model defaults from [scripts/orchestration/openclaw_cli_defaults.py](/Users/shenlan/workspaces/cloud-neutral-toolkit/github-org-cloud-neutral-toolkit/scripts/orchestration/openclaw_cli_defaults.py). Do not duplicate model selection logic inside prompts.
Do not duplicate role assignment logic outside `config/orchestration/roles.yaml`.

## Role Resolution

Resolve role intent before writing prompts:

- `codex_app`: chief engineer and acceptance gate
- `codex_cli`: direct executor for fallback-only work
- `opencode`: optional orchestration manager
- `antigravity`: optional automation manager
- `opencode_cli`: bounded editor worker
- `gemini_cli`: independent auditor worker
- `ollama`: cheap reviewer worker

Inspect the current role map with:

```bash
python3 scripts/orchestration/orchestration_roles.py
python3 scripts/orchestration/orchestration_roles.py codex_app
python3 scripts/orchestration/orchestration_roles.py --format json
scripts/orchestration/run_worker.sh --describe ollama
```

## Worker Selection

Use `opencode` for:

- bounded repo-aware edits
- playbook, workflow, and docs fixes
- low-risk implementation within explicit path boundaries

Use `ollama` for:

- read-only review
- cheap repeated checks
- config audits
- smoke-check prompt execution

Use `gemini` for:

- read-only cross-check
- short audit
- alternate diagnosis

## Prompt Contract

Every worker prompt must include four fields:

1. `Task`
2. `Allowed`
3. `Forbidden`
4. `Return only`

Keep prompts as short as possible while preserving those fields.

Default read-only prompt:

```text
Task:
- one concrete objective

Allowed:
- explicit read-only scope

Forbidden:
- secrets output
- unrelated edits
- production changes

Return only:
- finding
- blocker

From haitaopanhq/.github, Apache-2.0.

More from haitaopanhq/.github

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
Mar 13, 2026 (7 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
Apache-2.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from haitaopanhq

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get orchestration-run-task running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying orchestration-run-task?

CallGet help
Cite this page

orchestration-run-task skill. Agent Mart. https://agentmart.dev/skills/haitaopanhq-orchestration-run-task (last updated 2026-10-24)