ワンクリックでManusで任意のスキルを実行

始める

atlas-chunking

スター1

フォーク0

更新日2026年7月8日 17:21

Break a concept into chunks and generate questions to explore. Used by Atlas+Scout primitives.

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

JackReis

JackReis/hermes-skills

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

Atlas Chunking

Overview

Break a concept into chunks and generate questions to explore. First half of the Atlas+Scout compounding primitive (Nate's AI Meeting Takeaways).

Use Case: Prototype Nate's primitives:

Atlas: Chunk concepts (e.g., "source-of-truth discipline").
Scout: Prioritize chunks (e.g., "Is this worth working on?").

Input/Output

Input: A concept (e.g., "AI-enabled acquisitions"). Output: Chunks (subtopics) + questions (exploration prompts).

Procedure

Source: When the input says "latest posts" or another moving target, resolve the source first. Prefer an available canonical connector/API; if unavailable, use the public RSS/feed/archive and record the source URL, lookback window, selected item count, and access limitations in the JSON.
Decompose: Split each concept/post into 3–5 non-overlapping subtopics.
Question: For each subtopic, generate 2–3 questions to explore.
Validate: Ensure questions are open-ended and actionable.
Persist + verify: For cron/downstream use, write the requested JSON artifact, then validate it with a real JSON parser and structural checks before reporting completion.

Example

delegate_task(
    goal="Break 'AI-enabled acquisitions' into chunks and questions",
    context="Use the Atlas chunking pattern: subtopics + exploration questions.",
    toolsets=['terminal']
)

Verification

Chunks: 3–5 subtopics, no overlaps.
Questions: Open-ended, actionable.
Output: JSON format (for Scout compatibility).
Cron artifacts: See references/cron-json-pattern.md for moving-source cron jobs. The pattern documents source resolution metadata, access caveats, JSON structure, and validation steps for downstream automation.

このリポジトリの他の Skills

同じリポジトリ

cloudflare-one

JackReis/hermes-skills

Guides Cloudflare One Zero Trust and SASE work across Access, Gateway, WARP, Tunnel, Cloudflare WAN, DLP, CASB, device posture, and identity. Use when designing, configuring, troubleshooting, or reviewing Cloudflare One deployments. Retrieval-first: use current Cloudflare docs/API schemas instead of embedded product docs.

2026-07-161

cron-automation

JackReis/hermes-skills

Use when scheduling recurring polls, read-only monitors, or background jobs that must survive shell session boundaries.

2026-07-161

skill-library-management

JackReis/hermes-skills

Discover, register, and maintain Hermes skills from non-default paths (plugin directories, external repos, shared vaults). Use when skills are missing from `skills_list()`, when the user mentions plugin skills, when symlinking skills, or when reconciling duplicate/overlapping skills between the default tree and external sources.

2026-07-151

antigravity-cli

JackReis/hermes-skills

Operate the Antigravity CLI (agy): plugins, auth, sandbox.

2026-07-151

stream-retry-tuning

JackReis/hermes-skills

Tune HERMES_STREAM_RETRIES for mid-stream reconnect resilience.

2026-07-081

telegram-parallel-sessions

JackReis/hermes-skills

Manage parallel Hermes sessions in Telegram topics.

2026-07-081

name	atlas-chunking
description	Break a concept into chunks and generate questions to explore. Used by Atlas+Scout primitives.
version	1.0.0
author	Hermes
license	MIT
metadata	{"hermes":{"tags":["chunking","ideation","exploration"],"related_skills":["scout-prioritization","brainstorming"]}}

Atlas Chunking

Overview

Break a concept into chunks and generate questions to explore. First half of the Atlas+Scout compounding primitive (Nate's AI Meeting Takeaways).

Use Case: Prototype Nate's primitives:

Atlas: Chunk concepts (e.g., "source-of-truth discipline").
Scout: Prioritize chunks (e.g., "Is this worth working on?").

Input/Output

Input: A concept (e.g., "AI-enabled acquisitions"). Output: Chunks (subtopics) + questions (exploration prompts).

Procedure

Source: When the input says "latest posts" or another moving target, resolve the source first. Prefer an available canonical connector/API; if unavailable, use the public RSS/feed/archive and record the source URL, lookback window, selected item count, and access limitations in the JSON.
Decompose: Split each concept/post into 3–5 non-overlapping subtopics.
Question: For each subtopic, generate 2–3 questions to explore.
Validate: Ensure questions are open-ended and actionable.
Persist + verify: For cron/downstream use, write the requested JSON artifact, then validate it with a real JSON parser and structural checks before reporting completion.

Example

delegate_task(
    goal="Break 'AI-enabled acquisitions' into chunks and questions",
    context="Use the Atlas chunking pattern: subtopics + exploration questions.",
    toolsets=['terminal']
)

Verification

Chunks: 3–5 subtopics, no overlaps.
Questions: Open-ended, actionable.
Output: JSON format (for Scout compatibility).
Cron artifacts: See references/cron-json-pattern.md for moving-source cron jobs. The pattern documents source resolution metadata, access caveats, JSON structure, and validation steps for downstream automation.