Skip to content
Skills Library
AI & agents · Planning

subagent-dispatch

/subagent-dispatch

When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.

Repository
agentskel ↗
Licence
MIT
Last repository change
23.09.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 23.09.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.

Original author description

When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.

How it works

  1. When to Dispatch
  2. How to Dispatch
  3. Validate Results
  4. Rules
Original SKILL.md headings · 81 lines, 390 words
SKILL.md preview
---
name: subagent-dispatch
license: MIT
description: When delegating work to a subagent — implementation, review, research, or exploration. Use when a task benefits from a fresh context window, parallel execution, or isolated scope.
---

# Subagent Dispatch

**Principle:** The parent agent has full context (rules, memory, conventions). It
creates the plan. Subagents execute plan steps — they don't rediscover the project.

---

## When to Dispatch

| Signal | Why |
|---|---|
| Plan has independent steps that can run in parallel | Concurrent execution saves time |
| Context window is getting large | Fresh subagent avoids attention decay |
| Research would pollute main context with raw file reads | Subagent returns a summary |
| Code review needs a fresh perspective | Reviewer hasn't seen implementation reasoning |

**Do NOT dispatch when:** steps have sequential dependencies, or the task is simpler
to do directly than to explain.

---

## How to Dispatch

The parent creates the plan first (via task-planner or workflow Phase 1), then
identifies which steps can run in parallel.

### Implementation dispatch
Pass the specific plan steps. The parent already checked rules, conventions, and
sacred behaviors when creating the plan — the subagent doesn't need to re-read them.

```
"Execute steps 3-5 from the plan:
 - Step 3: [exact instruction]
 - Step 4: [exact instruction]
 - Step 5: [exact instruction]
 Files to modify: [list].
 Do not commit or push."
```

### Review dispatch
The reviewer needs to know what to check against. Include references to constraints.

```
"Review the changes made to [files].
 Read .memory/SACRED.md — verify no sacred behaviors were modified.
 Read .memory/CONVENTIONS.md — verify code follows project conventions.
 Check against the plan: [paste or summarize the plan].
 Report: issues found, verdict (approve/request changes)."
```

### Research dispatch
Pass the question. No rules needed.

```
"How does [module X] connect to [module Y]? Read MAP.md for orientation.
 Report: answer, evidence (file:line), relevant files."
```

---

## Validate Results

- [ ] Read the subagent's output
- [ ] Verify it meets the plan's success criteria
- [ ] Check for side effects (unexpected file changes)
- [ ] If validation fails, re-dispatch with corrected instructions

---

## Rules

1. **Subagents do NOT run session-start.** They execute plan steps, not session init.
2. **Subagents do NOT commit or push.** The parent handles git.
3. **One task per subagent.** Don't batch unrelated work.
4. **Only dispatch independent steps.** Sequential dependencies stay with the parent.

---

**No gate.** Dispatch is a judgment call based on task size and parallelism opportunity.

Versions

RepositoryScoreLicence
This page · ahmadulhoq/agentskel100MIT
wuyehaixww-hub/method-kit-v290MIT
chupeter854-lang/legal-agent-skills90MIT

Identical content appears in 1 other GitHub folders. We show one copy with the highest score.

Database snapshot: 13.11.2026 · package version 19172f34187b (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100subagent-driven-development
    Use when executing implementation plans with independent tasks in the current session. Triggers on: "execute the plan here", "dispatch subagents", "implement with reviews", any same-session execution request. Dispatches fresh subagent per task with two-stage review (spec…
  • 90claude-subagents
    Dispatch planning, review, implementation, or a scoped question to a fresh Claude subagent with the Agent tool — which subagent type and model to pick, what the prompt must carry, and how to check what comes back. Use whenever a workflow hands work to a Claude subagent,…
  • 100assembling-context
    Use when preparing context for subagent dispatch or managing token budgets. Triggers: 'prepare context for', 'assemble context', 'token budget', 'context package', 'what context does the subagent need'. Also invoked by develop during planning and execution phases.
  • 90subagent-delegation
    Dispatch to a subagent vs inline. TRIGGER when: full check suite or full-project verification; broad codebase search; first exploratory read; 2nd/3rd Bash toward same question; delegating implementation; delegating a multi-step gate/review loop. DO NOT TRIGGER when:…
  • 100common-subagent-driven-development
    Execute multi-task implementation plans by dispatching a fresh implementer subagent per task with independent review gates. Use when implementing plans, executing multi-task PRs, or transitioning from planning to implementation.
  • 90delegate
    Delegate a bounded task with /delegate and a self-contained brief. Use when work is independent, parallel, or needs a fresh context, and the parent must keep the scope, the diff review, and the checks.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.