|
18 | 18 | {"id":"seller-jzv","title":"1B: FreeWheel Data Models","description":"Pydantic models for FreeWheel entities parallel to gam.py (479 lines). Mapping: GAM Order = FW Insertion Order, GAM LineItem = FW Campaign+Placement, GAM AdUnit = FW Site\u003eSection\u003eAdSlot, GAM PrivateAuction = FW Programmatic Deal. Creates src/ad_seller/models/freewheel.py. Depends on 1A.","status":"open","priority":1,"issue_type":"feature","owner":"btomasette@gmail.com","created_at":"2026-02-26T18:01:32.169994-05:00","created_by":"Brian Tomasette","updated_at":"2026-02-26T18:01:32.169994-05:00","dependencies":[{"issue_id":"seller-jzv","depends_on_id":"seller-dcd","type":"blocks","created_at":"2026-02-26T18:02:08.121749-05:00","created_by":"Brian Tomasette"}]} |
19 | 19 | {"id":"seller-nur","title":"3B: Builder Guides for Vertical Customization","description":"# 3B: Builder Guides for Vertical Customization\n\n## Overview\nScenario-driven guides for teams customizing the seller agent for their specific business. Each guide walks through a real vertical with concrete examples of what to change and where.\n\n## Problem\nThe platform is highly configurable but it's not obvious how to:\n- Customize pricing models and discount tiers\n- Update prompts in sub-agents for domain-specific negotiation\n- Add business-specific \"skills\" (tools) to agents\n- Enforce compliance policies (HIPAA, political ad rules, etc.)\n- Configure inventory types and deal structures for different media\n\n## Planned Guides\n\n### Guide 1: Healthcare Publisher (e.g., WebMD)\n- HIPAA privacy skills that prevent prohibited targeting\n- Buyer vetting and restricted advertiser categories\n- Custom pricing for pharma vs. general advertisers\n- Compliance agent that validates proposals against regulations\n\n### Guide 2: CTV/Streaming Publisher (e.g., political ad rules)\n- Political advertising policies and seasonal overrides\n- Linear + CTV unified inventory setup\n- FreeWheel integration specifics\n- Content adjacency and brand safety rules\n\n### Guide 3: News/Media Publisher (general web)\n- High-volume programmatic with floor price management\n- Category blocking and sensitive content policies\n- Header bidding integration alongside direct deals\n- Seasonal pricing (elections, holidays, tentpole events)\n\n### Guide 4: Custom Pricing \u0026 Negotiation\n- How to modify pricing rules engine tiers and volume curves\n- Updating negotiation agent prompts for different strategies\n- Adding custom tools (e.g., competitive rate lookup, audience valuation)\n- Configuring approval gates for high-value deals\n\n## Each Guide Covers\n1. **Business context** — who this is for, what makes them different\n2. **Configuration changes** — settings.py, pricing tiers, feature flags\n3. **Agent customization** — prompt modifications, new tools, skill additions\n4. **Policy enforcement** — how to add compliance rules that can't be overridden\n5. **Testing** — how to validate customizations work correctly\n\n## Timing\nBest written after Phase 1 core is stable. Depends on having the full feature set available so guides aren't speculative.","status":"open","priority":2,"issue_type":"task","owner":"btomasette@gmail.com","created_at":"2026-02-27T16:03:34.622848-05:00","created_by":"Brian Tomasette","updated_at":"2026-02-27T16:03:34.622848-05:00","labels":["docs"]} |
20 | 20 | {"id":"seller-ryn","title":"3C: Deployment \u0026 Operations Guide","description":"# 3C: Deployment \u0026 Operations Guide\n\n## Overview\nProduction deployment and day-2 operations documentation. Pairs with 3G (IaC) to provide both the automation and the human-readable runbook.\n\n## Sections\n1. **Quickstart** — Local development setup, env vars, first API call\n2. **Deployment Options** — Docker, AWS (CloudFormation), Terraform, bare metal\n3. **Storage Configuration** — SQLite (dev) vs Redis (prod), migration path\n4. **Monitoring \u0026 Observability** — Event bus as audit log, key metrics, alerting\n5. **Scaling** — When to move from SQLite to Redis, horizontal scaling considerations\n6. **Security** — API authentication, buyer identity verification, secrets management\n7. **Troubleshooting** — Common issues, event bus diagnostics, approval timeout handling\n\n## Dependencies\n- Benefits from 3G (IaC) being complete for deployment automation sections\n- Benefits from 3A (API Reference) for cross-linking","status":"open","priority":3,"issue_type":"task","owner":"btomasette@gmail.com","created_at":"2026-02-27T16:03:36.612857-05:00","created_by":"Brian Tomasette","updated_at":"2026-02-27T16:03:36.612857-05:00","labels":["docs"],"dependencies":[{"issue_id":"seller-ryn","depends_on_id":"seller-299","type":"blocks","created_at":"2026-02-27T16:03:44.865149-05:00","created_by":"Brian Tomasette"}]} |
21 | | -{"id":"seller-tt5","title":"3A: API \u0026 Architecture Reference Documentation","description":"# 3A: API \u0026 Architecture Reference Documentation\n\n## Overview\nLiving documentation that gets updated incrementally as each feature lands. This is the technical reference for developers integrating with or extending the seller agent.\n\n## Sections\n1. **Architecture Overview** — System diagram, flow lifecycle, agent topology\n2. **REST API Reference** — All endpoints (products, pricing, proposals, deals, events, approvals), request/response schemas, error codes\n3. **Configuration Reference** — All settings.py options with defaults and env var names\n4. **Storage Schema** — Key patterns, data models, backend options (SQLite/Redis)\n5. **Event Bus \u0026 Approval Gates** — Event types, approval workflow, resume pattern\n6. **Agent \u0026 Crew Reference** — Each CrewAI agent's role, tools, and prompt structure\n7. **Extension Points** — Where to add custom tools, agents, pricing rules, ad server adapters\n\n## Approach\n- Created as a docs/ directory with markdown files\n- Updated incrementally as each Phase 1/2 feature lands\n- Each feature PR should include a docs update (even if just a new endpoint added to the API ref)\n\n## Deliverables\n- docs/architecture.md\n- docs/api-reference.md\n- docs/configuration.md\n- docs/storage.md\n- docs/events-and-approvals.md\n- docs/agents.md\n- docs/extending.md","status":"open","priority":2,"issue_type":"task","owner":"btomasette@gmail.com","created_at":"2026-02-27T16:03:16.080782-05:00","created_by":"Brian Tomasette","updated_at":"2026-02-27T16:03:16.080782-05:00","labels":["docs"]} |
| 21 | +{"id":"seller-tt5","title":"3A: API \u0026 Architecture Reference Documentation","description":"# 3A: API \u0026 Architecture Reference Documentation\n\n## Overview\nLiving documentation that gets updated incrementally as each feature lands. This is the technical reference for developers integrating with or extending the seller agent.\n\n## Sections\n1. **Architecture Overview** — System diagram, flow lifecycle, agent topology\n2. **REST API Reference** — All endpoints (products, pricing, proposals, deals, events, approvals), request/response schemas, error codes\n3. **Configuration Reference** — All settings.py options with defaults and env var names\n4. **Storage Schema** — Key patterns, data models, backend options (SQLite/Redis)\n5. **Event Bus \u0026 Approval Gates** — Event types, approval workflow, resume pattern\n6. **Agent \u0026 Crew Reference** — Each CrewAI agent's role, tools, and prompt structure\n7. **Extension Points** — Where to add custom tools, agents, pricing rules, ad server adapters\n\n## Approach\n- Created as a docs/ directory with markdown files\n- Updated incrementally as each Phase 1/2 feature lands\n- Each feature PR should include a docs update (even if just a new endpoint added to the API ref)\n\n## Deliverables\n- docs/architecture.md\n- docs/api-reference.md\n- docs/configuration.md\n- docs/storage.md\n- docs/events-and-approvals.md\n- docs/agents.md\n- docs/extending.md","status":"closed","priority":2,"issue_type":"task","owner":"btomasette@gmail.com","created_at":"2026-02-27T16:03:16.080782-05:00","created_by":"Brian Tomasette","updated_at":"2026-03-10T16:43:41.935742-04:00","closed_at":"2026-03-10T16:43:41.935742-04:00","close_reason":"Complete: 25+ seller doc pages, 15+ buyer doc pages, media kit docs, cross-linked GitHub Pages sites deployed","labels":["docs"]} |
22 | 22 | {"id":"seller-v8c","title":"1F: Multi-turn Session Persistence","description":"Persist conversation state across hours/days so buyer agents can resume negotiations. Session model with session_id, conversation_history, current_proposal, negotiation_state. Currently chat history is in-memory only. Modifies chat/main.py, storage/base.py, api/main.py. Creates src/ad_seller/models/session.py. No blockers.","status":"closed","priority":1,"issue_type":"feature","owner":"btomasette@gmail.com","created_at":"2026-02-26T18:01:01.779768-05:00","created_by":"Brian Tomasette","updated_at":"2026-02-28T14:00:09.562213-05:00","closed_at":"2026-02-28T14:00:09.562213-05:00","close_reason":"Closed"} |
23 | 23 | {"id":"seller-yu6","title":"1E: Linear TV Specialist Agent","description":"Full linear TV channel specialist for NBCU. Dayparts, scatter/upfront deals, commercial pod positions, GRP-based buying, audience guarantees. Products: NBC Primetime :30, Bravo Reality :15, Telemundo :30. Creates linear_tv_inventory_agent.py. Modifies flow_state.py (add linear_tv type), product_setup_flow.py (NBCU products), inventory_crews.py. Maps to PRD NBC inputs: GetProducts, GetTargeting, GetRate, GetAvails. No blockers.","status":"closed","priority":1,"issue_type":"feature","owner":"btomasette@gmail.com","created_at":"2026-02-26T18:01:00.307325-05:00","created_by":"Brian Tomasette","updated_at":"2026-03-04T09:55:41.908613-05:00","closed_at":"2026-03-04T09:55:41.908613-05:00","close_reason":"Implemented full Linear TV Specialist Agent: 5 Pydantic models (Daypart, MakegoodTerms, SupplyPoolEntry, LinearTVProduct, LinearDeal), 12 tools (pricing/avails/forecasting/traffic), Level 2 agent with institutional knowledge backstory, crew integration, PublisherCrew routing, 6 default products (3 NBCU direct + 2 Comcast MVPD + 1 reseller), 210 DMA codes, 58 unit tests, Rivian R2 cross-platform example. Standards: TIP v5.1.0, OpenRTB 2.6, IAB sellers.json."} |
0 commit comments