Skip to content
Skills Library
AI & agents · Planning

dispatching-parallel-agents

/dispatching-parallel-agents

Use when multiple frozen subtasks have non-overlapping write sets and can be implemented or reviewed in parallel.

Licence
Apache-2.0
Last repository change
04.08.2026

Assessment and security

90/ 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.
Content25 / 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 04.08.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Use when multiple frozen subtasks have non-overlapping write sets and can be implemented or reviewed in parallel.

Original author description

Use when multiple frozen subtasks have non-overlapping write sets and can be implemented or reviewed in parallel.

How it works

  1. Preconditions
  2. Fan-out Packet
  3. Model, Permission, And Token Boundaries
  4. Fan-in
Original SKILL.md headings · 33 lines, 266 words
SKILL.md preview
---
name: dispatching-parallel-agents
description: "Use when multiple frozen subtasks have non-overlapping write sets and can be implemented or reviewed in parallel."
---

# Dispatching Parallel Agents

## Preconditions

- Each subtask has goal, files, dependencies, material assumptions or explicit unknowns, acceptance checks, and verification.
- Write sets do not overlap.
- Shared foundation work is already done or assigned as a dependency.
- Fan-in owner is identified.

## Fan-out Packet

Each parallel subtask uses the six-block PM dispatch packet from `core-workflow-contract`. At minimum include:

- `task_intent`: `goal`, `task_type`, `work_mode`
- `frozen_scope`: `files_involved`, `write_set`, `dependencies`, `owner_lane`, `reviewer_lanes`, `review_package_ref`, `diff_ref`
- `execution_contract`: `assumptions`, `tradeoffs`, `simpler_option_considered`, `constraints`, `acceptance_checks`, `verification_budget`, `model_policy`, `model_cost_boundary`, `cost_boundary`, `reviewer_input_boundary`, `reviewer_permission_boundary`, `forbidden_approaches`, `read_before_write_targets` when code edits are planned
- `output_contract`: required skills, `required_artifacts`

Add `sub_task_id` for fan-in tracking.

## Model, Permission, And Token Boundaries

- Every parallel dispatch records the intended `owner_lane`, `reviewer_lane`, `permission_mode`, `write_set`, `review_package_ref` or `diff_ref`, `model_policy`, and `model_cost_boundary`.
- When the runtime supports subagent model selection, choose the cheapest adequate explicit model/tier for each lane instead of inheriting the current session default. If model selection is not enforceable, state that in `model_policy` and `model_cost_boundary.enforcement`.
- Batch fan-out that uses many reviewers, high-tier models, or repeated large context requires PM-visible cost acknowledgement or a cheaper explicit model/context strategy before dispatch.
- Review-only lanes use read-only permission mode and must not run mutating git/workspace commands. Reclassify to fix/implementation before allowing edits.
- Shared diffs, specs, logs, and review packages should be referenced once and reused by path or shard, not pasted into every subagent prompt.

## Fan-in

Adjudicate fan-in using the priority order from `core-workflow-contract` Fan-In Arbitration. Accept only structured handbacks with the required fields. Merge results by dependency order. Re-run integration verification after combining parallel work.

Versions

RepositoryScoreLicence
This page · funky-eyes/best-copilot90Apache-2.0
jjuraszek/pi-gauntlet100MIT
osmontero/opencode-skills100MIT
coco-research/coco100MIT
FrancoStino/opencode-skills-collection100MIT
seaworld008/commonly-used-high-value-skills100MIT
rootcastleco/rei-skills100MIT
nikospavlopoulos/devops-mutant-powers100MIT
Richicinschi/studio-os100MIT

Database snapshot: 13.11.2026 · package version 3e38f82922d6 (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

  • 100loom-dispatching-parallel-agents
    Dispatch multiple independent tasks to parallel subagents when no shared file conflicts exist. Use when: a confirmed plan has independent tasks that can run concurrently without touching shared files.
  • 100sr-dispatching-parallel-agents
    Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
  • 100orchestrate
    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.
  • 100parallel-execution
    Use when multiple subtasks have no shared files or dependencies and can be executed simultaneously.

Feedback and reports

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