writing-plans

Use when you have a spec or requirements for a multi-step task, before touching code

Install with Tessl CLI

npx tessl i github:obra/superpowers --skill writing-plans

What are skills?

Review — 57%

Does it follow best practices?

If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:

npx tessl skill review --optimize ./path/to/skill

Learn more

Evaluation — 86%

Agent success when using this skill

Validation — 11 / 11 Passed

Validation for skill structure

SKILL.md

Review

Evals

Evaluation results

100%

Create a Plan for Adding User Authentication

Plan document structure with correct header, save path, and announcement

Criteria

Without context

With context

Skill announcement

100%

Dated file path

100%

Header contains Goal

100%

Header contains Architecture and Tech Stack

100%

Sub-skill reference in header

100%

Without context: $0.1906 · 57s · 9 turns · 9 in / 3,160 out tokens

With context: $0.3857 · 1m 47s · 14 turns · 323 in / 7,137 out tokens

100%

95%

Create a Plan for Building a URL Shortener Library

TDD cycle with bite-sized steps and frequent commits

Criteria

Without context

With context

Write failing test step

100%

Verify test fails step

100%

Minimal implementation step

25%

100%

Verify test passes step

100%

Commit step per task

100%

Without context: $0.1662 · 40s · 8 turns · 57 in / 2,362 out tokens

With context: $0.3595 · 1m 23s · 16 turns · 582 in / 5,213 out tokens

65%

Create a Plan for Adding a Caching Layer

Execution handoff offering two choices with sub-skill references

Criteria

Without context

With context

Offers subagent-driven option

100%

Offers parallel session option

100%

References executing-plans sub-skill

60%

References subagent-driven-development sub-skill

Without context: $0.2576 · 1m 16s · 11 turns · 60 in / 4,349 out tokens

With context: $0.3248 · 1m 34s · 14 turns · 242 in / 5,001 out tokens

89%

19%

Create a Plan for Adding a Rate Limiter Middleware

Zero-context plan writing with complete task structure and numbered tasks

Criteria

Without context

With context

Numbered task format

40%

100%

Self-contained instructions

80%

88%

Multiple tasks covering distinct components

100%

Test file paths included

50%

75%

DRY/YAGNI adherence

80%

Without context: $0.1897 · 55s · 8 turns · 57 in / 3,130 out tokens

With context: $0.4069 · 1m 46s · 15 turns · 291 in / 7,206 out tokens

97%

64%

Create a Plan for Adding Email Notification Service

Exact file paths, complete code snippets, and runnable commands in tasks

Criteria

Without context

With context

Files section with exact paths

60%

100%

Complete code blocks

72%

88%

Exact run commands

100%

Expected output descriptions

100%

Without context: $0.2091 · 49s · 10 turns · 10 in / 2,956 out tokens

With context: $0.3513 · 1m 31s · 15 turns · 12 in / 5,906 out tokens

55%

35%

OAuth2 Single Sign-On Integration

@ syntax sub-skill references in plan steps

Criteria

Without context

With context

@ syntax used

Colon-separated namespace format

@ reference in header sub-skill callout

100%

Skill announcement

Dated plan file path

100%

Header: Goal, Architecture, Tech Stack

20%

100%

Files section with annotations

30%

100%

Complete code blocks

100%

Exact run commands

100%

Without context: $0.2382 · 1m 30s · 7 turns · 56 in / 5,410 out tokens

With context: $0.5819 · 2m 46s · 19 turns · 99 in / 10,602 out tokens

97%

44%

Add Test Coverage to Inventory Pricing Engine

Test design guidance for unfamiliar implementers

Criteria

Without context

With context

Specific test function signatures

100%

Concrete assert statements

66%

100%

Happy path and edge/error cases

100%

Expected failure output in TDD verify step

40%

70%

Exact pytest run command

50%

100%

Test file in Files section

100%

TDD step order maintained

30%

100%

Minimal implementation step

100%

Dated plan file path

100%

Plan header present

20%

100%

Without context: $0.2843 · 1m 30s · 9 turns · 58 in / 5,653 out tokens

With context: $0.6083 · 3m 7s · 24 turns · 21 in / 10,113 out tokens

85%

26%

Transactional Email Integration for User Notifications

Documentation references and zero-context completeness

Criteria

Without context

With context

External doc references present

75%

65%

Doc reference per relevant task

50%

Zero-context completeness

86%

80%

Complete code examples

90%

100%

Exact install/run commands

100%

Files section with annotations

50%

100%

TDD cycle steps

100%

Dated plan file path

100%

Plan header present

40%

100%

Sub-skill reference in header

100%

Without context: $0.3096 · 2m 3s · 7 turns · 56 in / 7,637 out tokens

With context: $0.7121 · 3m 22s · 24 turns · 298 in / 13,457 out tokens

Evaluated: 1 day ago
Agent: Claude Code

Table of Contents

Create a Plan for Adding User Authentication Create a Plan for Building a URL Shortener Library Create a Plan for Adding a Caching Layer Create a Plan for Adding a Rate Limiter Middleware Create a Plan for Adding Email Notification Service OAuth2 Single Sign-On Integration Add Test Coverage to Inventory Pricing Engine Transactional Email Integration for User Notifications

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.