Docs Work Items Rel R2 001 P0 Rel R2 001 QA Plan

Canonical source: docs/work-items/REL-R2-001/P0-REL-R2-001-QA-PLAN.md · Snapshot commit: 8739d34fda50

REL-R2-001 — task QA plan

• Task ID: REL-R2-001
• Artifact kind: qa
• Artifact state: draft
• Roadmap status: Backlog
• Milestone: R2
• Execution allowed: false
• Evidence boundary: Creation of this draft does not approve implementation, deployment, testing, restore, release, or production use.

QA inputs

• docs/product/releases/PRD-R2-TELEGRAM-PHOTO-CAPTURE.md
• docs/architecture/PHASE1-IMPLEMENTATION-PLAN.md
• docs/council/UX-DESIGN-REVIEW.md
• docs/design/UX-SPECIFICATION.md
• prototypes/calendar-ui/index-v5.html
• docs/council/execution/P0-PHASE1-TASK-DEFINITION-OF-READY.md
• docs/council/agents/P0-QA-LEAD.md
• docs/council/execution/P0-OWNER-ACTION-LEDGER.md

Test objective

Independently determine whether Media Privacy/Restore Acceptance satisfies its exact requirements and bounded claim in the named environment, using fictional/synthetic fixtures unless a later explicit human gate authorizes otherwise.

Scenario matrix

Scenario	Required coverage	Current state
REL-R2-001-QA-001	Happy path and exact task outcome: Execute capture, invalid input/date, album, duplicate, cover, Original, AI-exclusion, media restore, and rollback fixtures.	Draft
REL-R2-001-QA-002	Invalid, missing, duplicate, replayed, interrupted, timeout, stale, dependency-failure, and retry behavior	Draft
REL-R2-001-QA-003	Privacy/security/authorization, secret/log/cache/export/backup/evidence scans, and AI exclusions	Draft
REL-R2-001-QA-004	Schema, migration, compatibility, inventory, backup, separate-path restore, rollback/forward-fix	Draft
REL-R2-001-QA-005	Supported browsers, keyboard, focus, screen reader, contrast, 320 px, text/page zoom, landscape, themes, reduced motion	Draft
REL-R2-001-QA-006	Dependencies, co-resident/non-regression scope, performance/capacity bounds, observability and exact health states	Draft

Traceability and fixtures

• Requirement IDs: LID-TG-001, LID-TG-002, LID-TG-003, LID-TG-004, LID-TG-005, LID-TG-006, LID-TG-007, LID-TG-008, LID-TG-009, LID-TG-010, LID-OPS-005, LID-OPS-009, LID-OPS-011, LID-OPS-015, LID-OPS-018
• Dependencies: ENG-R2-001, ENG-R2-002
• Fixture class: fictional/synthetic; fingerprint and authentic-content exclusion required in executed evidence.
• Deferred negative scope: LID-UP-004, LID-DEF-001, LID-DEF-002, LID-DEF-003, LID-DEF-004, LID-DEF-005, LID-DEF-006 must remain absent.

Evidence bundle

Executed evidence records task/release/requirement/scenario IDs, source SHA, artifact and dependency/SBOM digests, fixture fingerprint, environment class, sanitized configuration digest, exact commands/tool versions, expected/actual results and timestamps, defects/retests, schema/migration state, backup/restore/rollback result, reviewer identity/independence, opaque private evidence reference when authorized, remaining limitations, and exact permitted claim.

Independence and severity gate

The executing QA reviewer must not be the candidate implementer. Any unresolved Sev-1/Sev-2, critical/high privacy or security finding, authentic-media violation, status/evidence mismatch, missing restore/rollback, optimistic Health behavior, missing authority, or specialist veto produces Hold or Roll back.

Independent QA disposition

Draft / Hold. QA is always required. Scenario design here is not executed evidence and cannot make the task Ready or Done.