ll-qa-e2e-from-proto

name	ll-qa-e2e-from-proto
description	ADR-053 unified entry skill for running the complete E2E test chain (proto→e2eplan→manifest→spec) with acceptance traceability. Orchestrates sub-skills via Skill tool.
version	1.0
adr	ADR-047, ADR-052, ADR-053
category	qa
chain	e2e-test
phase	full-chain
supersedes	[]

LL QA E2E From Proto

One-shot skill that runs the complete E2E test chain from a prototype or FEAT document.

Input

• prototype_freeze_package: directory path from ll-dev-feat-to-proto (proto mode)
• OR feat_ref: from ll-product-epic-to-feat (api-derived mode)

Output

• ssot/tests/e2e/{proto_id}/e2e-journey-plan.md — Contains Acceptance Traceability table
• ssot/tests/e2e/{proto_id}/e2e-coverage-manifest.yaml — Generated by ll-qa-e2e-manifest-init
• ssot/tests/e2e/{proto_id}/e2e-journey-spec/JOURNEY-*.md — Generated by ll-qa-e2e-spec-gen

Internal Pipeline

1. ll-qa-prototype-to-e2eplan — prototype/FEAT → e2e-journey-plan (with Acceptance Traceability)
2. ll-qa-e2e-manifest-init — e2e-journey-plan → e2e-coverage-manifest
3. ll-qa-e2e-spec-gen — e2e-coverage-manifest → e2e-journey-specs

Options

• --preview: Only run up to e2e-journey-plan, stop before manifest-init
• --no-spec: Stop after e2e-coverage-manifest, skip spec-gen
• --mode proto|api-derived: Force mode selection

Orchestrator Behavior

State Machine

START → E2EPLAN_RUNNING → TRACEABILITY_VALIDATING → MANIFEST_RUNNING → SPEC_RUNNING → COMPLETE

• START: Validate input (prototype OR feat_ref is provided, source is frozen)
• E2EPLAN_RUNNING: Call Skill tool /ll-qa-prototype-to-e2eplan with prototype_freeze_package or feat_ref
• TRACEABILITY_VALIDATING: Validate Acceptance Traceability table completeness
• MANIFEST_RUNNING: Call Skill tool /ll-qa-e2e-manifest-init with e2e-journey-plan output path
• SPEC_RUNNING: Call Skill tool /ll-qa-e2e-spec-gen with e2e-coverage-manifest output path
• COMPLETE: Return execution summary with paths to all generated artifacts

Error Handling Strategy

error_handling:
  strategy: fail-fast
  failure_at:
    e2eplan: abort
    manifest_init: abort
    spec_gen: continue  # preserve manifest on spec_gen failure
  artifacts_on_failure: keep
  retry: disabled

Acceptance Traceability Validation

validation:
  acceptance_traceability:
    check_mandatory_mapping: true     # each journey must have AC mapping
    check_complete_coverage: true   # each AC must have at least one journey coverage
    fail_on_incomplete: true
    validation_point: before_manifest_init

Non-Negotiable Rules

• Same as ll-qa-prototype-to-e2eplan (input validation, acceptance traceability)
• Same as ll-qa-e2e-manifest-init (cut rules, four-dimensional status)
• Same as ll-qa-e2e-spec-gen (spec completeness)
• Acceptance traceability table must be complete before proceeding to manifest-init
• Orchestrator does not directly generate artifacts — only calls sub-skills via Skill tool

Evidence

• Execution evidence: evidence/execution-evidence.schema.json
• Supervision evidence: evidence/supervision-evidence.schema.json