Development Using Claude Code or Codex · Part 7
Part 7: QA Execution Skill
April 13, 2026 · 1 min read
QaTestingRegression
QA should not be "run a few tests and hope." It should be structured, risk-based, and repeatable.
Skill for This Post: qa-execution
Create .ai/skills/qa-execution/SKILL.md:
---
name: qa-execution
description: Execute functional and regression QA on review-approved changes and produce release-quality bug reports.
---
# QA Execution
## Test Strategy
- Start with highest-risk user journeys
- Validate happy path, edge path, and failure path
- Include data validation, permissions, and state transitions
- Verify analytics/logging side effects when applicable
## Bug Report Format
For each bug include:
1) title
2) severity
3) steps to reproduce
4) expected result
5) actual result
6) environment
7) regression test recommendation
## Required Output
- passed scenarios
- failed scenarios
- bug list
- release recommendation: go | no-goPrompt Pattern
Run qa-execution on this branch.
Prioritize critical user journeys first.
Return bugs in the required format and a final go/no-go recommendation.Why This Skill Matters
- Makes QA outcomes auditable
- Forces clarity on release readiness
- Creates better inputs for UAT
Cross-Agent Install
ln -sfn "$PWD/.ai/skills/qa-execution" ~/.claude/skills/qa-execution
ln -sfn "$PWD/.ai/skills/qa-execution" ~/.codex/skills/qa-executionNext
Part 8 converts QA-verified builds into UAT scripts and stakeholder signoff.