ワンクリックでManusで任意のスキルを実行

delegation

スター3

フォーク0

更新日2026年7月4日 19:41

Load when writing a subagent brief for non-trivial dispatched work. Do not load for direct edits, trivial lookups, user-facing summaries, or delegation already covered by an existing brief.

インストール

Codex または Claude でインストールこの Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。

Manusで実行

ソース

AmbitiousRealism2025

AmbitiousRealism2025/Albion

GitHub リポジトリを開く Creator のリポジトリを見る

ダウンロード

Manusで実行

Delegation

Extend charter §5 with dispatch packets that constrain work and make returned evidence reviewable.

Brief Template

Section	One-line guidance
TASK	State the exact question or artifact; one active verb.
EXPECTED OUTCOME	Name the returned object, format, and maximum size.
CONTEXT	Provide only facts needed to work; include relevant paths.
MUST DO	List required checks, readings, and evidence.
MUST NOT DO	Fence scope, forbidden files, and prohibited actions.
TOOLS ALLOWED	Name read/write/network/tool limits explicitly.
SUCCESS CRITERIA	Define the termination condition and return shape.

Guard Rails

State a concrete termination criterion.
Fence scope with named paths, modules, or documents.
Forbid git state changes: no commits, checkouts, resets, rebases, stash, branch edits, or tag edits.
Require evidence in the return: file lines, command results, reproduction steps, or inspected artifacts.
Hide implementation narrative from verifier and counterexample-hunter; give task definition, artifacts, and expected behavior only.
Limit secrets exposure: name environment variables or secret locations, never values.
Require "not checked" labels for skipped checks with reasons.

Parallel Dispatch

Case	Rule
Independent reads	Dispatch in parallel; do not serialize.
Shared write target	Do not dispatch writes; split ownership or keep local.
Competing hypotheses	Assign one hypothesis per agent; require falsifier and result.
Verification	Use a fresh brief; omit the implementation path and your preferred explanation.

Reconciliation

Compare each return against its success criteria.
Mark evidence as observed, inferred, missing, or contradicted.
Resolve contradictions with the smallest direct check.
Update state-map.md, hypotheses.md, or verification.md before acting on the result.
Keep final judgment local; agents supply evidence, not authority.

このリポジトリの他の Skills

同じリポジトリ

hot-path

AmbitiousRealism2025/Albion

Load before implementing anything that runs continuously or per-item at scale — animation and render loops, timers, pollers, per-request middleware, large-collection iteration. Do not load for one-shot scripts, setup code, or paths that run a bounded handful of times.

2026-07-063

toolchain-fidelity

AmbitiousRealism2025/Albion

Load before hand-writing or hand-editing files a toolchain normally generates or owns — Xcode project files, lockfiles, generated schemas or migrations, IDE workspace files. Do not load for ordinary source code.

2026-07-063

maturity-assessment

AmbitiousRealism2025/Albion

Load when entering an unfamiliar codebase, or before proposing an architecture-level change. Do not load for tiny edits, known local fixes, post-implementation summaries, or tasks already scoped by a current state-map.md.

2026-07-053

recovery

AmbitiousRealism2025/Albion

Load on strike 3, a tangled run, or a contradiction that cannot be localized. Do not load for first failures, ordinary test red/green loops, or known environment outages.

2026-07-053

conductor

AmbitiousRealism2025/Albion

Load for cross-session orchestration — dispatching work to an Albion/GLM worker over tmux, polling a worker completion manifest, or reviewing worker output. Do not load for in-session subagent delegation; use delegation instead.

2026-07-043

completion-gate

AmbitiousRealism2025/Albion

Load when the stop gate blocks twice on the same task. Do not load for ordinary unfinished work, first gate blocks, or reports where no completion claim is being made.

2026-07-043

Delegation

Extend charter §5 with dispatch packets that constrain work and make returned evidence reviewable.

Brief Template

Section

One-line guidance

TASK

State the exact question or artifact; one active verb.

EXPECTED OUTCOME

Name the returned object, format, and maximum size.

CONTEXT

Provide only facts needed to work; include relevant paths.

MUST DO

List required checks, readings, and evidence.

MUST NOT DO

Fence scope, forbidden files, and prohibited actions.

TOOLS ALLOWED

Name read/write/network/tool limits explicitly.

SUCCESS CRITERIA

Define the termination condition and return shape.

Guard Rails

State a concrete termination criterion.

Fence scope with named paths, modules, or documents.

Forbid git state changes: no commits, checkouts, resets, rebases, stash, branch edits, or tag edits.

Require evidence in the return: file lines, command results, reproduction steps, or inspected artifacts.

Hide implementation narrative from verifier and counterexample-hunter; give task definition, artifacts, and expected behavior only.

Limit secrets exposure: name environment variables or secret locations, never values.

Require "not checked" labels for skipped checks with reasons.

Parallel Dispatch

Case

Rule

Independent reads

Dispatch in parallel; do not serialize.

Shared write target

Do not dispatch writes; split ownership or keep local.

Competing hypotheses

Assign one hypothesis per agent; require falsifier and result.

Verification

Use a fresh brief; omit the implementation path and your preferred explanation.

Reconciliation

Compare each return against its success criteria.

Mark evidence as observed, inferred, missing, or contradicted.

Resolve contradictions with the smallest direct check.

Update state-map.md, hypotheses.md, or verification.md before acting on the result.

Keep final judgment local; agents supply evidence, not authority.

name	delegation
description	Load when writing a subagent brief for non-trivial dispatched work. Do not load for direct edits, trivial lookups, user-facing summaries, or delegation already covered by an existing brief.