dispatch
Skill0 starsChanged 2 months agoMITView on GitHub
Dispatch a prompt to a background subagent and return control immediately. Use when the user says "/dispatch <prompt>", "run this in the background", "hand this off to a subagent", or wants to keep working while a task runs. Supports context-inheriting forks, fresh agents, worktree isolation, and model overrides.
40 other projects use a version of this.
Adopted by
Most recently taken up by
- renzrollon/interlocktheir own version, first seen Nov 5, 2026
- dan-myles/skillstheir own version, first seen Nov 1, 2026
- bestony/skillstheir own version, first seen Nov 1, 2026
- AcKeskin/contexturetheir own version, first seen Nov 1, 2026
- mhlaghari/stuntmantheir own version, first seen Nov 1, 2026
- vu1n/pillboxtheir own version, first seen Oct 26, 2026
Used with
Projects with it also use edge-detection, odds-analysis and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
noahowshSkilledge-detection
Comparing a tennis model's probability with the market: no-vig close, expected value at the best available price, closing line value, fractional Kelly sizing, and the checks that separate an estimated edge from noise. Use when the user has a probability and wants to know whether a price is better than fair, how large a stake is reasonable, or whether an edge is real. Requires a probability the user trusts. Do not use for fetching prices: see odds-explorer. Do not use for devig math alone: see odds-analysis. Do not use for historical testing: see backtesting.
Also in 4 projects that use this
- noahowshSkill
odds-analysis
Odds math for two-way tennis markets: American to implied probability, margin, devigging (multiplicative, power, Shin), no-vig fair odds, consensus across books, favorite-longshot bias, and expected value of a price. Use when the user asks what a price implies, how to remove the vig, what the true probability is, or how to compare books. Do not use for fetching or comparing live prices: see odds-explorer. Do not use for deciding a bet against a model: see edge-detection.
Also in 4 projects that use this
agiprolabsSkillwalk-forward-validation
Walk-forward validation framework for trading strategies and ML models with time-series-aware splits, overfit detection, and regime-aware validation
Also in 4 projects that use this
Yahmez3Skillbacktesting
Use when backtesting a trading strategy — evaluating a signal series against historical price data and computing performance metrics (Sharpe, Sortino, max drawdown, CAGR, win rate, expectancy). Triggers on "backtest this strategy", "what's the Sharpe on X", "run a walk-forward test", or "compare strategies A and B". Produces equity curves, trade logs, and tear sheets.
Also in 4 projects that use this
ffsshhttiikkSkillfeature-engineering
Machine learning feature engineering
Also in 4 projects that use this
varun29ankuSSkillorchestrate
This skill should be used when the user asks to 'orchestrate a task', 'break down work into parallel agents', 'coordinate subtasks', 'run agents in parallel', or mentions 'multi-agent'. Decomposes complex tasks into tracked subtasks, dispatches parallel subagents, and coordinates until completion.
Also in 4 projects that use this
- +3 more from noahowsh
Other versions40 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 40
- your-papa/obsidian-Smart2Brain★ 1.3ksame name
- s2b-dev/smart-second-brain★ 1.3ksame name
- bassimeledath/dispatch★ 411same name
- renzrollon/interlock★ 61same name
- fagemx/edda★ 36same name
- phj128/autoworker★ 18same name
- clonable-eden/cekernel★ 17same name
- mhlaghari/stuntman★ 16same name
- YippeeXuXu/dsh-collab-mode★ 5same name
- BigY0shi/super-skills★ 4same name
30 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- customer-facing-specSkill
- explain-step-by-stepSkill
- features-roadmapSkill
- markdown-to-formal-docSkill
- md-to-docxSkill
- plan-to-project-treeSkill
- review-planSkill
- runbookSkill
- writing-clearly-and-conciselySkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 38 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 22
No flags: pick sensibly — --fork if the prompt depends on session context, otherwise a fresh general-purpose agent.
- Very specific: line 39
If the prompt depends on session context but --fork was not given, use fork anyway (a fresh agent can't see the conversation).
- Very specific: line 41
Call the Agent tool once with description (3–5 words), the expanded prompt, and any flag-derived options.
- Distinctive: line 51
Never dispatch destructive or outward-facing actions (push, publish, delete, send) — those need the user watching.
- Distinctive: line 52
If already running as a subagent, refuse: execute inline instead of nesting.
“Common” is measured: kept by at least 6 in 10 of the 38 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Dispatch
Fire-and-forget: run the given prompt in a background subagent via the Agent tool, confirm launch in one line, and return control to the user. Results arrive later as a task notification.
Arguments
Everything after /dispatch is the prompt, minus leading flags:
| Flag | Effect |
--fork | subagent_type: "fork" — subagent inherits this conversation's full context. Use when the task refers to "this", "the PR we discussed", or anything established in-session. |
--worktree | isolation: "worktree" — force worktree isolation (default for code tasks, see below). |
--no-worktree | Opt out of worktree isolation; agent works directly in the current tree. |
--model <m> | Model override (sonnet, opus, haiku). Ignored with --fork. |
--type <t> | Explicit subagent_type (e.g. general-purpose, Explore, Plan). |
No flags: pick sensibly — --fork if the prompt depends on session context, otherwise a fresh general-purpose agent.
Worktree default for code tasks
A code task is any prompt that will create, edit, delete, or commit files in a git repository (fix, implement, refactor, migrate, write tests, apply review findings). For code tasks, pass isolation: "worktree" by default so the agent works on an isolated copy of the pertaining repo and never races the user's working tree. The worktree is created from the repo the task targets — if the prompt names a path in a different repo than the cwd, note that repo in the agent's prompt so it works there.
Skip worktree isolation (no flag needed) when:
- The task is read-only: review, explore, research, summarize, answer a question.
- The target is not inside a git repository.
- The user passed
--no-worktree.
--worktree still forces it on for anything else.
Procedure
- Parse flags off the front of the args; the rest is the task prompt.
- If the prompt depends on session context but
--forkwas not given, use fork anyway (a fresh agent can't see the conversation). - For a fresh (non-fork) agent, expand the prompt so it is self-contained: include file paths, PR numbers, branch names, and acceptance criteria the agent can't otherwise know. Tell it what to return, not what to display — its report comes back to you, not the user.
- Call the Agent tool once with
description(3–5 words), the expandedprompt, and any flag-derived options. Do not wait on it or poll. - Reply to the user with one line: what was dispatched and that results will arrive when it finishes. Then end the turn — do not block.
When results arrive
Relay the agent's findings to the user in your own words — the raw report is not shown to them. If the agent produced files or a branch, name the paths.
Constraints
- One dispatch per invocation. Multiple tasks = multiple
/dispatchcalls (they run concurrently). - Never dispatch destructive or outward-facing actions (push, publish, delete, send) — those need the user watching.
- If already running as a subagent, refuse: execute inline instead of nesting.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.