跳到正文
cccskills
免费在 GitHub 公开

platform-list

The Platform List: Arman's own register of large platform-level work he chose not to start yet. Use when the reminder hook says a row is due, when Arman says to add, snooze, kill, or finish a row, when you own a row, or when you think something belongs there. NOT for ordinary tasks, defects, or handoffs.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md5.1 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

<!-- SYNCED COPY — do not edit here. Canonical: common-docs/skills/platform-list/SKILL.md This file is distributed to every consuming repo by common-docs/meta/scripts/sync_skills.py. Edit the canonical, run the sync, and commit each repo. Edits made here are overwritten and lost. -->

The Platform List

One file: common-docs/operations/platform-list/LIST.md. One folder beside it with one short file per row. Owned by Arman. Never read it at session start. due.py beside it runs as a SessionStart hook on Arman's Mac and stays silent unless a Waiting row's Start date has arrived; it then tells the first attended session of the day, once, and no other. Read the list only when that reminder arrives or Arman asks about it.

What it is for

Arman normally starts big work the moment he decides on it. The rare exception is large, platform-wide work he has decided to do but not now. That is the whole list. It is not a backlog, not a defect ledger, not a place for handoffs, and not where your ideas go. Expect it to stay short. Almost every row starts from Arman's own words.

The four things you do

  1. Remind, only when the hook tells you to. Tell Arman in your first reply, in one plain sentence per row the reminder names. He will either say start it, or give a new date. A new date is a snooze: change the Start cell and nothing else. Never decide a date yourself.
  2. Suggest, never write. If you believe something belongs here, say so to Arman in one or two plain sentences inside your normal reply. You do not add rows, files, drafts, or notes anywhere. He adds it or he does not.
  3. Add only on his word, in this conversation. When Arman names an item to add, write one row and one item file, both from the templates below. A new row has an empty Owner note and an empty Inbox. The Item is one sentence under fifteen words that reads as a vision, not a task.
  4. Own a row honestly. When you claim a row, set Status to Started, fill Owner with your template line, and keep the Owner note true. Clear Inbox lines once you have read them. When you stop for the day, the row must be readable cold by the next owner.

Status, exactly

Waiting nobody started · Started an owner is working · Blocked stopped on something the Owner note names · Live shipped and in use but could still slip away unfinished. No Done exists.

Deleting a row

A row and its file are deleted, together, in exactly two cases:

  • Arman says he is killing it.
  • The work has reached the point where it is impossible for it to not be finished. Sometimes that is ten minutes in, because the code now depends on it. Sometimes that is weeks after it went live, when the last proof lands. Arman judges this, and you ask him when you think a row has reached it.

Deletion is complete: the row, the file, and nothing else. No "deleted" line, no archive folder, no history file, no copy in a plan or handoff or register, no summary of what was here. The bundle log says only that the Platform List changed, never which item. Building any place that remembers what was on this list is forbidden, whoever asks. check.py in the folder enforces the shape; the docs lint runs it.

Owner line, one per tool

Fill Owner so a stranger can find your conversation. Use the exact form for your tool:

ToolOwner cell
Claude CodeClaude Code · <session title> · <session id> · <machine or cloud> (session id is the UUID of your transcript under ~/.claude/projects/; resume with claude --resume <id>)
CodexCodex · <thread title> · <thread id> (resume with codex resume <id>)
CursorCursor · <chat title> · <workspace folder> · <YYYY-MM-DD started>
Anything else<tool> · <title> · <the id its resume command takes>

Templates

Row: | [<item, under 15 words>](<slug>.md) | Waiting | <YYYY-MM-DD> | | | |

Item file <slug>.md, at most 250 words of body, these five headings and no others:

---
type: Platform Item
title: "<item>"
description: "<one sentence>"
timestamp: <YYYY-MM-DD>T00:00:00Z
---

## What this is
## Why it matters
## What finished looks like
## Where it stands
## Decisions

Write plain sentences. No jargon, no codenames, no lists of files. The code, commit messages, and the system's own docs carry detail; this file carries only what someone needs to pick the work up and understand where it stands. When it grows past the cap, cut, never split.

Scope test

Not written yet. Arman will shape it with the first real rows. Until then the test is his word.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Admin debug system wiring for the floating AdminIndicator's live debug data and Copy Full Context. Use when adding debug visibility to a route or feature, wiring useDebugContext or the debug panel, capturing console errors, or enabling the copy-context workflow for a page.

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

Compact two-icon Copy / Copy-for-AI controls (components/agent-copy). Use when adding copy buttons to a row, card, list, or record; merging duplicate Copy/JSON/AI controls; continuing the copy rollout; or writing a Copy-for-AI payload. NOT for markdown content actions (use rich-document-actions).

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

Disclosing a surface's existing fixed AI jobs in the shell's top Agents menu. Use when a page, panel, overlay, or window already runs a mandate behind a button, assist, automatic action, or mode; when the agent-disclosure guard names a file; or during a surface check.

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

The watch-fix-rerun method for making an existing platform agent or automated process efficient and correct: baseline its ledger, run one unit yourself, fix the class behind every wasted call, rerun, record. Use when asked to improve, optimize, watch, or 'make efficient' an agent, a sandbox session, a sync job, or any recurring automated process.

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

Redux state for agent execution and firing agent shortcuts. Use when editing features/agents/redux/, building agent UI, creating a conversation, touching assembleRequest or NDJSON stream state, wiring agent-state selectors, adding a per-conversation capability, or triggering a shortcut from a button, menu, or code (useShortcutTrigger, launchAgentExecution).

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

The provision for a call site: the exhaustive menu of values that place in the code can realistically produce. Use before creating or fixing a mandate or agent that reasons about more than its own input, when deciding whether an agent can answer at all with what it is sent, or when its output 'looks right' but nobody checked it could be. NOT for building the agent (use create-agent).

中文简介正在准备中。当前显示原文说明。

armanisadeghi/ai-matrx32026年10月22日 更新

查看 armanisadeghi 的全部技能

报告此技能的问题