clarify

Runs a four-stage clarification funnel with critic review before surfacing user decisions.

Updated Sep 14, 2026
One-click install
npx skills add https://github.com/pandejesal/drone-nav-sar --skill clarify-pandejesal
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: clarify
Source: https://github.com/pandejesal/drone-nav-sar/tree/main/.swarm/bundled-skills/clarify
Command: npx skills add https://github.com/pandejesal/drone-nav-sar --skill clarify-pandejesal

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Ambiguous user requests lead to plans built on unstated assumptions or excessive clarifying questions. This protocol inventories every material uncertainty, resolves what it can from context and a critic sounding board, and surfaces only genuine user decisions as a structured packet. ## Core Features & Use Cases - Four-Stage Clarification Funnel: Inventories all material uncertainties, classifies each as self_resolved, critic_resolved, research_needed, user_decision, or deferred_nonblocking, then consults a critic sounding board before asking the user anything. - Always-Surface Protection: Categories like scope boundaries, destructive behavior, security risk, and breaking changes must reach the user; DROP verdicts on these items are overridden to ASK_USER. - Assumption Recording: Every resolved uncertainty is recorded as an explicit assumption in the spec, plan, or context file, preventing silently dropped decisions. - Use Case: When asked to plan a feature with unclear migration and rollout implications, the architect runs this funnel to resolve defaults internally and presents only the blocking scope and compatibility decisions to the user. ## Quick Start Activate MODE: CLARIFY on an ambiguous request to run the clarification funnel and produce a structured user decision packet.

Frequently Asked Questions about clarify

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How does the clarification funnel decide which questions to ask the user?

The funnel inventories all material uncertainties, classifies each into five categories, then sends research_needed and user_decision items to a critic sounding board. Only items the critic approves as genuinely requiring user input are surfaced in the final decision packet.

What is the critic sounding board in the clarify protocol?

The critic sounding board reviews candidate clarification questions and returns a verdict: UNNECESSARY, RESOLVE, REPHRASE, or APPROVED. Verdicts map to funnel actions that drop, resolve, reword, or escalate each item before the user is interrupted.

Which uncertainty categories must always be surfaced to the user?

Always-surface categories include scope boundaries, data loss, security and privacy risk, backward compatibility, breaking changes, new dependencies, deprecations, cross-platform impact, cost tradeoffs, UX choices, rollout strategy, and QA gate strictness. The critic cannot drop these items.

What happens when research for an uncertainty takes too long?

Research_needed items have a fixed 5-minute protocol budget. If research does not complete in time, the item is automatically reclassified as user_decision with a note that research was incomplete, preventing the funnel from stalling.

How are resolved uncertainties prevented from being silently dropped?

Every item classified as self_resolved, critic_resolved, or deferred_nonblocking must be recorded as an explicit assumption in the spec, plan, or .swarm/context.md. Silently dropping any uncertainty that entered the funnel is defined as a protocol violation.