ingest-spec（摄取规格）

Ingest Spec

When to use

• After analyze-codebase, when you have a spec to parse
• When running as part of run-testing-session pipeline (Stage 2)
• When adding new scenarios to an existing test plan

Inputs

• Spec content: a .feature file path or pasted plain-English test cases
• (Optional) docs/playwright-spec-testing/project-context.md for context

What it does

Parse raw spec input into structured scenarios and save to docs/playwright-spec-testing/parsed-spec.md.

Phase 1: Identify input format

Gherkin has Feature:, Scenario:, Given/When/Then keywords. Plain English has numbered steps, "Expected:" sections, or prose descriptions.

Phase 2: Parse into scenarios

For each scenario, extract:

• Name: The scenario title
• Preconditions: Setup state (from Background/Given or Prerequisites)
• Steps: Ordered actions (from When/And or numbered steps)
• Expected outcomes: Assertions (from Then or Expected section)
• URL hints: Any URL path mentioned (e.g., /login, /dashboard)

For plain-English input, map prose to this structure. If something is ambiguous, preserve it verbatim in a "Notes" field — do NOT infer steps not in the spec.

Phase 3: Write parsed-spec.md

Append to (or create) docs/playwright-spec-testing/parsed-spec.md:

# Test Plan

_Last updated: [DATE]_

---

## Scenario: [Name]

**Source:** [file path or "pasted input"]
**Format:** [Gherkin / Plain English]

### Preconditions
- [condition]

### Steps
1. [step]

### Expected Outcomes
- [outcome]

### URL Hints
- [url]

### Test File
_To be assigned by plan-tests_

### Exploration Report
_To be generated by explore-app_

### Status
- [ ] Planned
- [ ] Explored
- [ ] Generated
- [ ] Passing

Add one block per scenario. If parsed-spec.md already exists, append — do not overwrite.

Key Rules

• Never infer steps not in the spec
• Preserve all URLs verbatim
• Do not merge multiple scenarios into one
• Do not discard information

Output

docs/playwright-spec-testing/parsed-spec.md

Report when done:

• Status: DONE | DONE_WITH_CONCERNS | BLOCKED | NEEDS_CONTEXT
• Number of scenarios parsed
• List of scenario names
• Any ambiguities noted