spec-format-experiment
Run A/B experiment comparing GWT vs EARS+properties as spec formats for AI agent implementers. Generates specs in both formats for 5 test features, dispatches implementers, scores results. Use when evaluating spec format effectiveness for the pipeline-v3 spec writer.
npx skills add majiayu000/claude-skill-registry --skill spec-format-experiment --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Spec Format Experiment: GWT vs EARS+Properties > **Note:** This skill is a **manual experiment protocol**, not automated tooling. The operator runs each step manually by dispatching subagents with the reference prompts and recording results. Future iteration: could be automated with an orchestrator skill. ## Purpose Compare two behavioral spec formats for AI coding agents: - **GWT** (Given-When-Then): Scenario-based, verbose, explicit test data - **EARS+Properties** (Easy Approach to Requirements Syntax + universal invariants): Requirement-based, compact, invariant-focused Both formats sit on top of the same **contract-as-code** (compiled C# DTOs). Only the behavioral spec format varies. ## Hypothesis - **H0:** GWT and EARS+properties produce equivalent results - **H1:** EARS+properties is more token-efficient with equal or better completeness and implementer success ## Commands | Argument | Action | |----------|--------| | `run-specs` | Generate specs for all 5 features in both formats (3 runs each = 30 total) | | `run-specs feature-N` | Generate specs for feature N only | | `run-impl` | Dispatch backend-implementer for all primary specs (2 trials each = 20 total) | | `run-impl
- Purpose
- Hypothesis
- Commands
- Protocol
- Phase 1: Generate Specs
- Phase 2: Score Specs
- Phase 3: Implement From Specs
- Phase 4: Verdict
- Minimum Viable Experiment
- Token Budget
- Bias Mitigation
- Reference Files
What does the spec-format-experiment skill do?
Run A/B experiment comparing GWT vs EARS+properties as spec formats for AI agent implementers. Generates specs in both formats for 5 test features, dispatches implementers, scores results. Use when evaluating spec format effectiveness for the pipeline-v3 spec writer.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill spec-format-experiment --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.