Skip to content

strategy-testability-review

An independent, adversarial reviewer (test-engineer persona) that assesses whether a refined strategy can be validated. Invoked by strategy-review in an isolated context: fork, it auto-detects local vs CI mode and checks whether acceptance criteria are testable and measurable, what edge cases are missing, what the test strategy would be, whether NFRs are quantified with numeric thresholds, and whether every threshold cites a grounded source (RFE, architecture doc, or SME input) rather than being invented. It emits a structured per-strategy verdict with untestable criteria and missing edge cases listed.

Plugin: strat-creator | Internal

Contract

Skill Contract canonical-skill-v1

Review strategy features for testability — measurable acceptance criteria, edge case coverage, and validation feasibility.

Identity

Functions
review
Success
  • Produces a testability assessment for each strategy with recommendation.
  • Identifies untestable criteria and missing edge cases.

Optimization Targets

task_success judge SKILL.md @ 0e102d4

Invariants

Must Preserve
  • Do not approve strategies with unmeasurable acceptance criteria.
  • Do not invent test scenarios unsupported by the strategy content.
Fixed Context
toolsRead, Grep, Glob
knowledgerepository_contentpublic, task_inputtask_private

Traceability

Skill
.claude/skills/strategy-testability-review/SKILL.md

Diagram

strategy-testability-review diagram

Arguments

/strategy-testability-review [RHAISTRAT-NNNN]
Argument Required Default Description
RHAISTRAT-NNNN - A strategy key to review only that strategy. If omitted, all strategies in the tasks directory are reviewed.

Usage

/strategy-testability-review RHAISTRAT-1531