一键在 Manus 中运行任何 Skill

para-second-brain

星标3

分支1

更新时间2026年2月19日 21:19

Use this skill when the user wants to organize, classify, or maintain a PARA-method second brain. Triggers include asking where to file something, distinguishing projects from areas, processing an inbox, setting up a new project, completing or archiving a project, running a monthly review, validating system structure, or finding stale/orphaned content.

安装

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

在 Manus 中运行

来源

robdefeo

robdefeo/agent-skills

打开 GitHub 仓库查看创作者相关仓库

下载

在 Manus 中运行

PARA Method

Use this skill to help users organize and maintain a second brain using the PARA system (Projects, Areas, Resources, Archives).

Routing

Pick the entry point based on user intent:

Classification and "where does this go?" questions: read references/decision-trees.md
Example requests and edge-case comparisons: read references/examples.md
Operational process requests (inbox, review, setup, close-out, archive): read references/workflows.md
Troubleshooting pain points and validation guidance: read references/common-problems.md

If the request is broad or does not clearly match one route, default to references/decision-trees.md.

Output Convention

Classification guidance and Q&A: answer in chat
Validation workflows: run scripts/validate.sh and write report output to PARA-validation-YYYY-MM-DD.md in the current working directory
Installation location: out of scope for this skill; installation is handled by separate tooling

Terminology

Use "second brain" for the user's vault/folder structure
Use "PARA system" only for the method/framework

Validation Workflow

When the user asks to validate structure or project health:

Read references/common-problems.md for interpretation guidance.
Run scripts/validate.sh <path> (or omit path to use current directory).
Save report output to PARA-validation-YYYY-MM-DD.md if user wants a file.
Summarize critical findings and recommended next actions.

同仓库更多 Skills

同仓库

scratchpad

robdefeo/agent-skills

Capture out-of-scope discoveries, ideas, tasks, and questions into a per-repo SCRATCHPAD.md while coding, then retrieve, archive, or promote them later. Use this skill whenever the user wants to write something down to deal with later instead of now — "park this", "note that", "jot this down", "add a note", "look at later", "I'll lose this otherwise". Route these here even when it looks like a trivial inline note — the point is durable, retrievable capture, not answering in the moment. Also use it to read items back ("what's open", "what did I park", "any notes about X"), tick them done, archive, or promote a parked item to a GitHub issue, memory, or project. And fire it yourself when, mid-task, you hit something outside the current scope — often unexpected behavior to work around but not fix now — that would otherwise be lost. This is for TRANSIENT, project-local capture; it is NOT durable memory (facts auto-loaded every session), reminders, meeting notes, PR text, or code comments.

2026-06-203

skill-scope

robdefeo/agent-skills

Decide how to scope a Claude skill before building or restructuring it: what belongs in one skill, what splits off, what merges, whether the work needs a new skill at all. Use BEFORE skill-creator whenever the user is (1) creating a new skill, (2) significantly restructuring or expanding an existing skill, (3) splitting one skill into two or merging two into one, or (4) unsure whether a workflow needs a new skill vs. extending an existing one. Skill-creator should not start scaffolding until skill-scope has run. Fires on: "create a skill", "make a skill", "new skill", "design a skill", "build a skill", "scope this skill", "should this be one skill or two", "is this skill too big", "is this skill too narrow", "where does this skill belong", "split this skill", "merge these skills", restructuring an existing skill, or `/skill-scope`. Skip for trivial edits where scope is settled (typos, trigger phrase tweaks, body rewording, packaging) — these don't reopen scope decisions.

2026-06-073

pr-address-feedback

robdefeo/agent-skills

Use this skill whenever the user wants to address, respond to, work through, or handle GitHub PR review comments — including code review feedback, reviewer suggestions, bot comments, or requested changes — even if they don't say "address". Handles the full end-to-end workflow: fetch threads, plan fixes, commit, push, and reply. Requires `gh` CLI.

2026-04-303

pr-create

robdefeo/agent-skills

Use this skill whenever the user wants to create, open, or raise a new GitHub pull request from the current branch — including phrasings like "open a PR", "raise a PR", "make a PR", "ship this". Drafts a Conventional Commits title and a Summary body that explains *why* and *what*, respects `.github/pull_request_template.md` when present, auto-links referenced issues, waits for approval, then pushes and opens the PR. Assumes commits already exist. Requires `gh` CLI.

2026-04-303

diataxis-documentation-guidance

robdefeo/agent-skills

Use this skill to classify, scope, split, draft, and review documentation with the Diataxis framework. Triggers include choosing tutorial/how-to/reference/explanation mode, resolving mixed-mode pages, rewriting docs into a target mode, and validating drafts with a checklist.

2026-03-063

amazon-working-backwards

robdefeo/agent-skills

Guide the Amazon Working Backwards process from the 5 Questions through to a full PR-FAQ document. Use when the user wants to (1) draft answers to the Amazon 5 Questions for a proposal or idea, (2) refine or critique existing 5Q answers, (3) verify and challenge 5Q answers with probing questions, (4) generate a PR-FAQ document from 5Q answers, (5) review or critique an existing PR-FAQ against Amazon standards, or (6) clarify any element of the 5Q or PR-FAQ process. Triggers include mentions of "5 questions," "PR-FAQ," "PRFAQ," "working backwards," "press release FAQ," Amazon-style proposals, or requests to write/review product proposals in Amazon format.

2026-02-163

name	para-second-brain
description	Use this skill when the user wants to organize, classify, or maintain a PARA-method second brain. Triggers include asking where to file something, distinguishing projects from areas, processing an inbox, setting up a new project, completing or archiving a project, running a monthly review, validating system structure, or finding stale/orphaned content.

PARA Method

Use this skill to help users organize and maintain a second brain using the PARA system (Projects, Areas, Resources, Archives).

Routing

Pick the entry point based on user intent:

Classification and "where does this go?" questions: read references/decision-trees.md
Example requests and edge-case comparisons: read references/examples.md
Operational process requests (inbox, review, setup, close-out, archive): read references/workflows.md
Troubleshooting pain points and validation guidance: read references/common-problems.md

If the request is broad or does not clearly match one route, default to references/decision-trees.md.

Output Convention

Classification guidance and Q&A: answer in chat
Validation workflows: run scripts/validate.sh and write report output to PARA-validation-YYYY-MM-DD.md in the current working directory
Installation location: out of scope for this skill; installation is handled by separate tooling

Terminology

Use "second brain" for the user's vault/folder structure
Use "PARA system" only for the method/framework

Validation Workflow

When the user asks to validate structure or project health:

Read references/common-problems.md for interpretation guidance.
Run scripts/validate.sh <path> (or omit path to use current directory).
Save report output to PARA-validation-YYYY-MM-DD.md if user wants a file.
Summarize critical findings and recommended next actions.