Claude Code Skills

Claude Code Skill

Test Plan Writer

Use when writing a test plan or test strategy for a release, feature, sprint or project, defining QA scope, risks, test types, environments, entry and exit criteria, or estimating testing effort.

Install

npx skills add WST2002/claude-code-skills --skill test-plan-writer

Or copy skills/test-plan-writer into ~/.claude/skills/.

Overview

A test plan decides where testing effort goes and when the product is good enough to ship. Drive it by risk: test hardest what would hurt users or the business most if it broke.

Method

  1. Understand the release: features and changes (PRD, tickets, changelog, diff), affected areas, integrations, deadlines, team size.
  2. Assess risk for each area: impact if it breaks × likelihood (new code, complex logic, past bugs, many integrations). Rank areas high, medium, low.
  3. Choose test types per area based on risk: functional, regression, integration, API, end-to-end, performance, security, accessibility, compatibility, exploratory.
  4. Define environments and data: where tests run, browsers and devices, test accounts, data setup and reset.
  5. Set entry and exit criteria that can be checked objectively.
  6. Estimate and schedule effort per area; note what is automated vs manual.
  7. Write using the format below; keep it to what the team will actually use.

Format

# Test Plan — <release / feature>

1. Scope
   In scope: <features, platforms>
   Out of scope: <and why>

2. Risk assessment
   | Area | Impact | Likelihood | Risk | Test approach |

3. Test types and coverage
   | Type | Areas | Manual / Automated | Owner |

4. Environments and test data
   <environments, browsers/devices matrix, accounts, data setup>

5. Entry criteria
   e.g. build deployed to staging, smoke tests pass, test data ready

6. Exit criteria
   e.g. all P1 cases pass, no open Blocker/Critical bugs, Major bugs triaged,
   regression suite ≥ 95% pass, performance within agreed limits

7. Schedule and effort
   | Activity | Estimate | Dates |

8. Defect management
   <tool, severity definitions, triage cadence>

9. Risks to the plan and mitigations
   <late builds, unstable environments, missing data>

10. Deliverables
   <test cases, automation results, test summary report>

Common mistakes

  • Equal effort on every area instead of focusing on risk.
  • Exit criteria like "testing complete" that nobody can measure.
  • No out-of-scope section, so expectations differ.
  • Copying a generic template that does not mention the actual features.
  • Forgetting regression of untouched areas that share code with the change.

View source on GitHub

Related skills