Start here → root
AGENTS.md(agent entry). This file is the router (ai-docs/SPEC_INDEX.md); system overview inARCHITECTURE.md. LoadAGENTS.md+ this file first; pull every other doc on demand. Context-efficiency: link to canonical docs — don't duplicate them; route to the minimum needed per task.
AI agent entry point after
AGENTS.md. Load this once at session start; pull other docs on demand. Source of truth:.sdd/manifest.json(this file mirrors it for humans).
| Module | Responsibility | Manifest coverage state | Start here |
|---|---|---|---|
store/ |
MobX singleton; global CC state; proxies SDK events; sole SDK access point | DRAFT | packages/contact-center/store/ai-docs/store-spec.md |
cc-components/ |
Shared presentational React UI primitives | DRAFT | packages/contact-center/cc-components/ai-docs/cc-components-spec.md |
cc-widgets/ |
r2wc Web Component wrappers (aggregator) | DRAFT | packages/contact-center/cc-widgets/ai-docs/cc-widgets-spec.md |
cc-digital-channels/ |
Digital channels (chat/email/social) widget | DRAFT | packages/contact-center/cc-digital-channels/ai-docs/cc-digital-channels-spec.md |
station-login/ |
Agent login: team + device selection | DRAFT | packages/contact-center/station-login/ai-docs/station-login-spec.md |
user-state/ |
Agent state: state, idle codes, timer | DRAFT | packages/contact-center/user-state/ai-docs/user-state-spec.md |
task/ |
Task widget bundle: CallControl, IncomingTask, OutdialCall, TaskList, CallControlCAD | DRAFT | packages/contact-center/task/ai-docs/task-spec.md |
ui-logging/ |
Metrics/telemetry: withMetrics, metricsLogger |
DRAFT | packages/contact-center/ui-logging/ai-docs/ui-logging-spec.md |
test-fixtures/ |
Shared test mocks/helpers | DRAFT | packages/contact-center/test-fixtures/ai-docs/test-fixtures-spec.md |
@webex/widgets/ |
Legacy meetings widgets (separate family) | DRAFT | packages/@webex/widgets/ai-docs/widgets-spec.md |
| If the task is… | Load |
|---|---|
| Understanding the system | ARCHITECTURE.md |
| Working in a module | that module's <module>-spec.md (see registry); load only the relevant section |
| Changing the store / SDK access | store-spec.md + ARCHITECTURE.md Component Interaction; check @webex/contact-center types (node_modules/@webex/contact-center/dist/types/index.d.ts) |
| Adding/changing a public surface (export, custom element, event) | CONTRACTS.md first, then the owning module spec |
| A UI/component change | cc-components-spec.md + patterns/react-patterns.md |
| A new widget | new-widget templates (ai-docs/templates/new-widget/); create the module spec as part of the change |
| A bug fix | the touched module spec + patterns/ + RULES.md; bug-fix template ai-docs/templates/existing-widget/bug-fix.md |
| Playwright E2E work | playwright/ suites + ai-docs/templates/playwright/ |
| Naming a concept | GLOSSARY.md |
| Anything touching input/identity/logging | SECURITY.md |
| Review & merge | REVIEW_CHECKLIST.md |
What kind of change?
├─ New feature / widget -> create/update an intake record under .generated/sdd/features/<KEY>/run-records/
├─ Bug / defect -> create/update an intake record; load the touched module spec
├─ New module / component -> create the module spec (module-spec template) + register in .sdd/manifest.json
└─ Doc/spec backfill only -> no feature intake record; reconcile the target spec, regenerate, run conformance
The intake record confirms scope/modules against the code and sets the change class (routine / security / contract / perf-critical / ui) that gates conditional spec sections.
| INC id | Date | Module | One-line | Link |
|---|---|---|---|---|
| none recorded | — | — | Populate from the tracker as RCAs are written | — |
| Doc | Location | Purpose |
|---|---|---|
| Agent entry | AGENTS.md (root) |
First file read: commands, critical rules, boundaries, routing |
| Architecture | ai-docs/ARCHITECTURE.md |
System components, interactions, monorepo package map |
| Patterns | ai-docs/patterns/ |
Repo conventions (TypeScript, React, MobX, testing), correct vs incorrect |
| Rules | ai-docs/RULES.md + ai-docs/rules/ |
Enforceable do/don't beyond AGENTS critical rules |
| Glossary | ai-docs/GLOSSARY.md |
Ubiquitous language: term → definition → code location |
| Security | ai-docs/SECURITY.md |
Trust boundaries, secret/token handling, data classification |
| Contracts | ai-docs/CONTRACTS.md |
Root index of public surfaces (exports, custom elements, events) |
| Service state | ai-docs/SERVICE_STATE.md |
Living as-built registry — read first to avoid duplicating a surface |
| Getting started | ai-docs/GETTING_STARTED.md |
Clone/build/test loop, workspace layout |
| Decision records | ai-docs/adr/ |
Standing ADRs — why the architecture is the way it is |
| Review catalog | ai-docs/REVIEW_CHECKLIST.md |
6-core + 4-coverage + 3-cross-cutting review checks |
| SDK reference | @webex/contact-center types (node_modules/@webex/contact-center/dist/types/index.d.ts) |
installed SDK .d.ts surface — verify every SDK call |
No DATA_MODEL.md: this repo owns no persistent datastore (all domain data comes from the SDK at runtime).