AD SEO Agent decomp
internal prototype · canonical JSON + Dreamborn Forge HTML
internal generated
decomp · supabase_json

AD SEO Agent decomp

decomp artifact · for AD SEO Agent · status draft

Planning Surface

Use this to decide what happens next.

Status

draft

Agent Handoff
Start Here

decomp artifact · for AD SEO Agent · status draft

Completion Evidence

No explicit evidence field yet. Require tests, screenshots, linked PRs, or reviewed outputs before marking complete.

Artifact Shape
  • gates: 4 items
  • status: draft
  • handoff: blocking_module_ids: object, notes_for_plan_compiler: object, ready_for_task_planning: boolean
  • modules: 9 items
  • project: id: string, name: string, client_id: string, platform_project_id: string, platform_project_title: string
  • dev studio: views: object, render_as: string
  • model reviews: 4 items
  • pre plan audit: id: string, ref: string, status: string, summary: string, blockers: object, ready_for_plan: boolean, resolved_decisions: object
  • schema version: 1.0
  • dependency graph: root_modules: object, critical_path: object, parallel_groups: object, terminal_modules: object
  • source artifacts: policy_ref: object, ux_spec_ref: string, research_ref: object, m00_audit_ref: string, architecture_ref: string, requirements_ref: string, coherence_review_ref: string
  • decomposition strategy: basis: string, rationale: string, parallelism_strategy: string, context_window_strategy: string
Structured Payload

Machine-readable source fields

gates
idnametypereturn pathafter modulesevidence required
GATE-001Host Alignment Gatereviewerdecomposition- M-00- client project route files and existing partial edits inventoried. - /seo confirmed as canonical implementation target. - Standalone prototype marked context-only.
GATE-002Data Contract Gateautomatedrequirements- M-01- Fixture validation passes. - At least 10 recommendations and five agent roles exist. - Seed data includes source mode, ROI assumptions, evidence, action, governance tier, and white-label context. - Bundled fixture mode verified without Supabase credentials. - Supabase import/mirror path exists or is explicitly deferred without blocking Monday runtime. - Demo-local approval state interface exists.
GATE-003Approval Demo-Local Implementation Gateexecdecomposition- M-04- ApprovalGate service abstraction uses sessionStorage-backed demo-local state. - Approval interactions make no network calls when persistence is disabled. - If Justin reopens persistence before build, architecture/decomposition returns before dispatching M-04.
GATE-004Monday Production Demo Gateexecdecomposition- M-08- Build/test results. - Desktop and mobile browser QA screenshots. - No blocking console errors. - Seeded fallback path verified with provider credentials absent. - Cloudflare route/deploy readiness confirmed.
status

draft

handoff
blocking module ids

No items captured.

notes for plan compiler
  • M-00 is complete. PLAN may be created, but build dispatch still requires Justin explicit approval.
  • Apply M-00 audit classifications to task ownership: app/page.tsx is route composition prototype, components/seo-studio.tsx is replace/refactor, data fixture is refactor, app/vendor/blanc is committed Monday release snapshot from upstream Blanc.
  • AD SEO Agent is a client project. Use C:/Users/jstnk/Downloads/cc/projects/seo-studio; do not use bezeliq-studio, Dev Studio, or internal Studio repos.
  • Do not dispatch build tasks until Justin explicitly approves task planning/build execution.
  • Keep M-01 and M-02 as the first planning slice; downstream UI work should depend on their contracts.
  • M-02 owns app/page.tsx as the route/layout composer. M-03, M-04, and M-05 own child components only and must not edit the route file in parallel unless the final PLAN serializes those edits.
  • Approval persistence is an accepted risk and must remain separable from the read-only Monday demo path.
  • PLAN.json should translate these modules into final task packets, worker ownership, dependency gates, verification commands, and artifact refs.
  • Monday data path is bundled static Hubbard fixture import. Supabase Projects/Demo seed/migration work is required as an importable mirror but must not block the first read-only route unless Justin/Vikram explicitly switch runtime source mode.
  • ApprovalGate defaults to demo-local/read-only service abstraction. It must not issue a network call to API-004 unless approval persistence is explicitly enabled by the exec gate.
  • Approval persistence decision is answered for Monday: demo-local/read-only sessionStorage. Do not add API-004 persistence tasks unless Justin explicitly reopens the decision before build dispatch.
  • Route composition is its own serialized module M-07; parallel child component modules must not edit app/page.tsx directly.
  • Supabase Projects/Demo instance is fyauumztocuinwhcokbn; do not target RedKey platform Supabase for AD SEO Agent product/demo seed data.
ready for task planning

true

modules
idgoalnamerisksscopestatus
M-00Confirm C:/Users/jstnk/Downloads/cc/projects/seo-studio as the client repo, audit its premature implementation against the artifacts, and define remediation before PLAN.json.Host Confirmation and Prototype Audit- RISK-M00-001: Agents may implement in the standalone prototype or internal Dev Studio instead of client project.- in: 5 items - out: 4 items - handoff to other modules: 1 itemcomplete
M-01Create the shared data contract and complete Hubbard seed records that power the read-only production pilot and future Supabase Projects/Demo import.AI Object Model and Hubbard Seed Dataset- RISK-M01-001: Weak seed data will make an otherwise polished UI feel fake.- in: 7 items - out: 4 items - handoff to other modules: 1 item
M-02Integrate AD SEO Agent as a client project at /seo with optional /products/seo redirect and a responsive Blanc-based product shell.client project Product Route and Blanc Shell- RISK-M02-001: Route shell may accidentally inherit internal Dev Studio assumptions.- in: 6 items - out: 3 items - handoff to other modules: 1 item
M-03Build the first buyer-facing product moment: Hubbard identity, ROI before rankings, ranked recommendations, evidence inspection, and agent activity.Command Center and Opportunity Experience- RISK-M03-001: UI could look beautiful but still sell a normal SEO dashboard instead of an AI-first operating surface.- in: 8 items - out: 4 items - handoff to other modules: 2 items
M-04Implement governed action tiers and ApprovalGate behavior while isolating write persistence behind the deferred auth/policy decision.Governance and Approval Flow- RISK-M04-001: Approval persistence could expand scope into auth/RLS work and endanger Monday demo.- in: 7 items - out: 5 items - handoff to other modules: 1 item
M-05Add release-critical System/Data status for seeded source mode, provider availability, freshness, and fallback behavior.System/Data Status- RISK-M05-001: White-label section could read like marketing instead of product proof.- in: 3 items - out: 3 items - handoff to other modules: 1 item
M-06Add the affiliate distributor white-label proof needed to sell network packaging without overbuilding admin.White-Label Story- RISK-M06-001: White-label section could read like marketing instead of product proof.- in: 3 items - out: 4 items - handoff to other modules: 1 item
M-07Wire M-03, M-04, M-05, and optional M-06 child components into app/page.tsx so /seo renders the complete product route.Route Composition Integration- RISK-M07-001: Parallel child component work may not integrate cleanly into the route shell.- in: 4 items - out: 3 items - handoff to other modules: 1 item
M-08Verify the AD SEO Agent route is production-ready on desktop/mobile, has no blocking console/build failures, and can be released on Cloudflare for the Monday pilot.Production QA, Cloudflare Release, and Monday Demo Gate- RISK-M06-001: A visually strong local route may still fail production due to Cloudflare routing, build, or environment differences.- in: 5 items - out: 3 items - handoff to other modules: 1 item
project
id

ad-seo-agent

name

AD SEO Agent for Affiliated Distributor Networks

client id

ad

platform project id

0.0.8832348

platform project title

AD SEO Agent

dev studio
views
  • module_board
  • dependency_graph
  • parallel_groups
  • critical_path
  • module_scope
  • gates
  • risks
  • open_questions
render as

decomposition

model reviews
idmodelstatusprovidertask typeevidence ref
MODEL-DECOMP-001sonnetappliedclaudearchitecture-reviewtmp/model-runsad-seo-agent-claude-architecture-review-ask-model.json
MODEL-DECOMP-002sonnetappliedclaudearchitecture-reviewtmp/model-runsad-seo-agent-claude-architecture-review-after-fixes.json
MODEL-DECOMP-003sonnetappliedclaudearchitecture-reviewtmp/model-runsad-seo-agent-claude-architecture-review-final.json
MODEL-M00-001sonnetappliedclaudearchitecture-reviewtmp/ad-seo-agent-m00-audit-claude-review-3.json
pre plan audit
id

M00-AUDIT-001

ref

docs/projects/ad-seo-agent/artifacts/M00_AUDIT.json

status

complete

summary

C:/Users/jstnk/Downloads/cc/projects/seo-studio is confirmed as the canonical AD SEO Agent client repo. Existing implementation is unapproved prototype output and must be remediated against artifacts before PLAN/build.

blockers

No items captured.

ready for plan

true

resolved decisions
blanc

For Monday Cloudflare Pages, use committed app/vendor/blanc release snapshot; upstream source of truth remains C:/Users/jstnk/Downloads/cc/projects/blanc. Do not depend on sibling file package in production until packaging/monorepo hardening exists.

route

Canonical route https://bezeliq.ai/seo. Optional /products/seo alias belongs to parent bezeliq.ai Cloudflare/domain routing layer.

model review

Claude M-00 architecture review completed in tmp/ad-seo-agent-m00-audit-claude-review-3.json and findings applied.

verification

seo-studio lint, typecheck, and build pass after hygiene/config fixes.

schema version

1.0

dependency graph
root modules
  • M-00
critical path
  • M-00
  • M-01
  • M-02
  • M-03
  • M-07
  • M-08
parallel groups
idreasonmodule ids
PG-001Seed/data contract and route shell can be prepared together after host alignment; both must complete before M-03.- M-01 - M-02
PG-002Core surfaces, governance UI, System/Data status, and white-label preview share the M-01 data contract and M-02 shell. M-06 is P1 and may be cut without blocking P0 QA.- M-03 - M-04 - M-05 - M-06
terminal modules
  • M-08
source artifacts
policy ref

No value captured.

ux spec ref

docs/projects/ad-seo-agent/artifacts/UX_SPEC.json

research ref

No value captured.

m00 audit ref

docs/projects/ad-seo-agent/artifacts/M00_AUDIT.json

architecture ref

docs/projects/ad-seo-agent/artifacts/ARCHITECTURE.json

requirements ref

docs/projects/ad-seo-agent/artifacts/REQUIREMENTS.json

coherence review ref

docs/projects/ad-seo-agent/artifacts/COHERENCE_REVIEW.json

decomposition strategy
basis

mixed

rationale

Decompose by the smallest sequence that can ship a production-ready Monday pilot: data/object foundation, client repo/route integration, AI-first product surfaces, governance behavior, white-label/status polish, then production verification and deploy. The split follows approved feature, UX surface, and architecture boundaries while keeping the approval persistence risk isolated from the read-only demo path. Monday runtime data is explicitly the bundled Hubbard fixture import; Supabase Projects/Demo receives importable/mirrored seed data but is not a runtime blocker for the Monday route.

parallelism strategy

Foundation and route integration form the critical start. Once seed/object shapes and the route shell exist, UI surface modules can run in parallel against the same fixture contract. Governance persistence can run separately behind a decision gate, while QA/deploy waits for product surfaces to converge.

context window strategy

Each module owns a bounded slice with explicit source refs and approximate paths. A planner can load only the module, the cited artifact sections, and the target repo files for that slice instead of the full artifact set.