deep-plan

Structured clarification before decisions. Use when user is in PLANNING mode, explicitly asks to plan or discuss, or when agent faces choices requiring user input. Ensures agent asks questions instead of making autonomous decisions when multiple valid approaches exist or context is missing.

1.12x

Quality

76%

Does it follow best practices?

Impact

56%

1.12x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/deep-plan/SKILL.md

Evaluation results

40%

-1%

Planning a New Data Pipeline Service

Question format and phased clarification

Criteria

Without context

With context

Context questions first

50%

41%

Priority emoji used

Blocking questions present

Category headers

100%

Confidence indicators

Options with rationale

40%

Question limit respected

Context question coverage

100%

Phase delineation

37%

87%

No autonomous decisions

100%

56%

10%

Mixed Backlog: Planning Session vs. Quick Fixes

Activation discrimination and priority categorization

Criteria

Without context

With context

Typo fix: no questions

100%

Variable rename: no questions

100%

Payment item: questions raised

100%

Notification item: questions raised

100%

Context questions included

100%

Priority emoji categorization

Blocking questions identified

Question count limit

50%

100%

Confidence indicators

Category section headers

20%

Options with reasons

33%

74%

10%

Implementing a User Role System

Mid-execution re-check and decision log consent

Criteria

Without context

With context

Ambiguity explicitly flagged

100%

Stop signal at ambiguity

33%

66%

No silent assumption

83%

100%

Clarification questions for ambiguity

100%

Recommendation with confirmation request

50%

80%

Priority marker on ambiguity

Confidence indicator present

No unsolicited decision file

100%

Unanswered question revisited

37%

50%

Clear decided vs. open sections

100%

Repository: idrevnii/perks
Commit: 51dd9ad

Evaluated: 25 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Planning a New Data Pipeline Service Mixed Backlog: Planning Session vs. Quick Fixes Implementing a User Role System

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.