choplin/my-agent-skills/discuss-toolkit-one-point

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Controls the pacing of a discussion with multiple known open points. Applies when two or more points are about to be presented, or when the user asks to proceed one at a time. Shows the whole set as a compact labeled list, then resolves or parks exactly one point before opening the next.

Install

npx vskill@latest install choplin/my-agent-skills/discuss-toolkit-one-point

SKILL.md

one-point - One Discussion Point at a Time

Keep a complex discussion navigable by refusing to present many discussion points at once. Summarize the whole landscape as a list, then drill into one point at a time.

Why This Skill Exists

In high-abstraction discussions (design, architecture, planning), the AI tends to surface every open point it notices and throw them all at the user in a single burst. The user cannot engage with all of them at once — the discussion becomes impossible to follow and forward progress stalls. This skill enforces the discipline of one point at a time so each point gets real attention and gets resolved before the next opens.

Definition — "discussion point" (論点): a single open question, decision, or trade-off that needs the user's input to move forward. Two facets of the same decision are one point; two independent decisions are two points.

Invocation

1. Explicit user invocation (primary)

The user calls /one-point (or says "一つずつ", "論点を整理して", "一気に出さないで") — usually as a course-correction right after the AI presented several points at once. Treat it as: "stop, give me the map, then take these one at a time."

2. AI autonomous application (secondary)

Before presenting your reply, check: am I about to raise two or more discussion points? If yes, apply this skill's procedure instead of dumping them. Prevent the burst rather than waiting to be corrected.

Do not over-fire: a single point, a clarification interview (that is dig), or a quick A-vs-B with clear context do not need this.

Procedure

Step 1 — Show the whole landscape as a list

Enumerate every open point as a short labeled list — one brief label plus a one-line gloss each. No full detail yet. This gives the user the whole map without demanding they engage all of it at once.

論点は 3 つあります:
1. **認証方式** — セッション or トークンか
2. **エラー時の挙動** — 失敗をどこまでユーザーに見せるか
3. **保存先** — DB か外部ストレージか
まず論点1から詰めます。
Step 2 — Pick which point to open first
  • If order matters (one point blocks or reframes another), open the most

foundational / blocking one first, and say why.

  • If order is free, let the user choose which to tackle first.
Step 3 — Present exactly ONE point in detail

For the current point only:

  • Restate the point and the context that makes it a real decision.
  • Lay out the concrete options / trade-offs.
  • Give your recommendation with reasoning (do not hide behind a neutral menu —

take a position).

If you use a structured question tool (e.g. AskUserQuestion), it must cover this one point only, even though the tool can hold several questions. Multiple facets of the same point are fine; a second, distinct point is not.

Step 4 — Resolve or park, then move on
  • Fully resolve the current point (user decides/confirms) or explicitly

park it (agree to defer, and note what is still open) before opening the next.

  • Re-show the remaining list as you advance so the user keeps the map:

"論点1は解決。残りは 2, 3。次は論点2。"

Anti-patterns

  • Dumping — presenting all points with full detail in one message. This is

the exact behavior this skill exists to stop.

  • Batching distinct points into one question — one AskUserQuestion (or one

paragraph) that asks about several unrelated decisions at once.

  • Silent term introduction — opening a point built on a concept you have not

defined. Define the term first (with one concrete example), then use it.

  • Opening the next point before closing the current one — leaving a trail of

half-discussed points.

Relationship to Other Skills

  • dig — clarifies or broadens thinking while tracking the whole discussion.

one-point assumes the discussion is framed enough; it manages the pacing of multiple known points. Independent skills.

Success Criteria

  • [ ] All open points were shown as a compact labeled list before any detail.
  • [ ] Only one point was presented in detail at a time.
  • [ ] Each point was resolved or explicitly parked before the next was opened.
  • [ ] The remaining-points list was visible as the discussion advanced.
  • [ ] No structured question mixed two distinct points in one round.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/choplin/my-agent-skills/discuss-toolkit-one-point/badge)](https://verified-skill.com/skills/choplin/my-agent-skills/discuss-toolkit-one-point)

Something wrong with this skill or its page? Tell us.