From cc55984dd132454b1cdd0cbe520c56937e47f863 Mon Sep 17 00:00:00 2001 From: citizen204 Date: Thu, 25 Jun 2026 04:18:25 +0930 Subject: [PATCH] fix(brainstorming): suppress AskUserQuestion; add explicit conversational-text rule MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On v6.0.0 the platform-neutral bootstrap reframe caused Claude Code to map "ask clarifying questions / present multiple-choice options" to the native AskUserQuestion tool, overriding the skill's conversational intent. Add an explicit note in brainstorming/SKILL.md and in the Claude Code tool mapping reference that asking the user is not a tool action on Claude Code — questions must be presented as plain conversational text in the terminal. Fixes #1773 --- skills/brainstorming/SKILL.md | 1 + skills/using-superpowers/references/claude-code-tools.md | 1 + 2 files changed, 2 insertions(+) diff --git a/skills/brainstorming/SKILL.md b/skills/brainstorming/SKILL.md index b0d52b2589..fa69c61c8c 100644 --- a/skills/brainstorming/SKILL.md +++ b/skills/brainstorming/SKILL.md @@ -70,6 +70,7 @@ digraph brainstorming { - For appropriately-scoped projects, ask questions one at a time to refine the idea - Prefer multiple choice questions when possible, but open-ended is fine too - Only one question per message - if a topic needs more exploration, break it into multiple questions +- **Present questions as plain conversational text — do not use structured-choice tools (e.g. `AskUserQuestion`).** Ask in the chat/terminal interface directly. - Focus on understanding: purpose, constraints, success criteria **Exploring approaches:** diff --git a/skills/using-superpowers/references/claude-code-tools.md b/skills/using-superpowers/references/claude-code-tools.md index 7ddd549aa7..20c56c496f 100644 --- a/skills/using-superpowers/references/claude-code-tools.md +++ b/skills/using-superpowers/references/claude-code-tools.md @@ -19,6 +19,7 @@ Skills speak in actions ("dispatch a subagent", "create a todo", "read a file"). | Multiple parallel dispatches | Multiple `Agent` calls in one response | | Task tracking ("create a todo", "mark complete") | `TaskCreate`, `TaskUpdate`, `TaskList`, `TaskGet`; `TodoWrite` in `claude -p` / Agent SDK unless `CLAUDE_CODE_ENABLE_TASKS=1` is set | | Background-process / subagent lifecycle (read output, cancel) | `TaskOutput`, `TaskStop` — these are distinct from the todo tools above and apply to running shells, agents, and remote sessions | +| **Ask the user a question / present multiple-choice options** | **Plain conversational text in the terminal — do NOT use `AskUserQuestion`.** Skills that say "ask one question at a time" or "prefer multiple choice" mean to type the question/options directly into the chat, not invoke a structured-choice tool. | ## Instructions file