A Claude Code skill that enables structured, heuristic-driven exploratory testing sessions with automatic MCOASTER report generation.
Invoke /exploratory-testing with a test charter and Claude will:
- Execute an exploratory testing session guided by industry-standard heuristics
- Automatically produce a full MCOASTER report at the end — no prompting needed
- Log every bug found with ID, severity, criterion, evidence, and a suggested fix
- Deliver a prioritised recommendations list
| Mnemonic | Purpose |
|---|---|
| FEW HICCUPPS | Test oracles — how to recognise a problem |
| SFDIPOT | Product coverage — what areas to explore |
| MCOASTER | Session reporting structure |
| I SLICED UP FUN | Mobile app testing |
| FCC CUTS VIDS | Application touring strategies |
| RCRCRC | Regression test prioritisation |
| FAILURE | Error handling evaluation |
| RIMGEA | Bug advocacy and reporting |
| CRUSSPIC STMPL | Non-functional quality attributes |
| + more | Data attack cheat sheet, data type heuristics, wisdom rules |
/plugin install exploratory-testing@danashby/Exploratory-Testing-Skill
/exploratory-testing "Explore the search screen, with WCAG 2.2 AA in mind, to discover accessibility risks"
Use the charter format:
"Explore [target], with [resources/constraints], to discover [information about risks]"
Every session produces:
- MCOASTER report sections (Mission, Coverage, Obstacles, Audience, Status, Techniques, Environment, Risks)
- Numbered bug list (BUG-01, BUG-02...) with severity, criterion, evidence, and fix
- Summary table
- Prioritised recommendations
MIT