skills/copilotkit__copilotkit
1
0
Fork 0
mirror of https://github.com/CopilotKit/CopilotKit.git synced 2026-09-14 16:26:20 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
codex/cloudplot-showcase-migration
copilotkit__copilotkit/showcase/integrations/google-adk/qa/subagents.md
T
Tyler Slaton ef1ca9808c feat(showcase/google-adk): QA-blitz parity ports — agent prompts, tool surface, e2e specs 
Result of 10 parallel QA agents auditing all 30 active demos against
langgraph-python (north-star). Each agent ported drift back to LP-verbatim
across three axes:

1. Agent layer
   - tool_rendering_common.py: rebuilt to LP's surface — get_weather,
     search_flights(origin, destination), get_stock_price, roll_d20,
     roll_dice. Removed the ADK-only query_data.
   - tool_rendering_*_agent.py (4 variants): ported LP's travel/concierge
     prompt; reasoning-chain variant got LP's chain-two-tools prompt.
   - beautiful_chat_agent.py: ported LP's per-tool system prompt; added
     manage_sales_todos / get_sales_todos / generate_a2ui; dropped the
     redundant schedule_meeting (frontend HITL handles it).
   - open_gen_ui_agents.py: ported LP's full SYSTEM_PROMPT for both
     variants, including the Websandbox.connection.remote.* contract
     for the advanced sandbox demo (was `window.sandbox.*`, which the
     LP frontend's Websandbox bridge silently no-ops).
   - byoc_agents.py: fused LP's hashbrown + json-render prompts so the
     single ADK byoc_agent emits both wire shapes. Aliases exported for
     a future per-route split.
   - declarative_gen_ui_agent.py: ported LP's a2ui_dynamic SYSTEM_PROMPT.
   - a2ui_fixed_agent.py: picked up LP's #4734 regression guard
     ("exactly ONCE", "do NOT call again").
   - agent_config_agent.py: rewrote to read useAgentContext (was
     state["config"]); reconciled schema to LP's 3-field camelCase
     {tone, expertise, responseLength} with LP's value enums.
   - subagents_agent.py: dropped the "running" placeholder; returns
     plain str so the LP-verbatim frontend's `result?.trim()` works.
   - hitl_in_app_agent.py / hitl_in_chat_book_call_agent.py: prompts +
     tool-result shape ({approved, reason}) aligned to LP.
   - AGUIToolset() added wherever it was missing on the bespoke agents
     (multimodal, mcp_apps, a2ui_fixed) so frontend-registered tools
     reach the model.

2. Dedicated runtime routes
   - copilotkit-multimodal/route.ts (new) — mirrors LP shape with
     ADK's HttpAgent + AGENT_URL pattern.
   - copilotkit-agent-config/route.ts (new) — same pattern.
   - copilotkit-mcp-apps/route.ts — refreshed.

3. Frontend ports (ADK frontend brought to LP-verbatim where it had
   drifted from the parity blitz state)
   - tool-rendering family (4 demos): full re-port — WeatherCard,
     FlightListCard, StockCard, D20Card, ReasoningBlock, CatchallRenderer,
     suggestions, and the page wiring with all useRenderTool /
     useDefaultRenderTool / reasoningMessage registrations.
   - a2ui-fixed-schema, mcp-apps, multimodal: full frontend re-ports
     with their _components/ Tailwind primitives.
   - frontend-tools, frontend-tools-async, agent-config: ported LP's
     component structure (separate Background, NotesCard with query_notes,
     config-context-relay).
   - shared-state-read, shared-state-read-write, readonly-state-agent-context:
     ported LP's demo-layout + _components + suggestions. recipe-card.tsx
     pulled directly from LP (one QA agent had adapted to Unicode glyphs
     thinking ADK lacked lucide-react — it doesn't, after the parity blitz).
   - shared-state-streaming, subagents, hitl-in-app: ported LP's
     DocumentView / supervisor-activity / TicketsPanel structure.
     hitl-in-app/page.tsx pulled directly from LP to keep the hyphenated
     agent slug aligned with the renamed registry key.
   - auth, hitl-in-chat: ported LP's SignInCard-first auth UX and the
     time-picker Tailwind port.
   - prebuilt-popup: pulled LP's main-content + suggestions split.

4. Test fixtures
   - 30 tests/e2e/<slug>.spec.ts ported from LP, several overwriting
     stale stubs (shared-state-streaming, subagents, auth, hitl-in-chat,
     shared-state-read, agent-config).
   - 30 qa/<slug>.md ported from LP with ADK env-var and registry
     references substituted (GOOGLE_API_KEY, AGENT_URL, registry.py).
   - QA3's byoc-hashbrown / byoc-json-render specs renamed to
     declarative-hashbrown / declarative-json-render with internal
     URL references substituted (the orchestrator pass had already
     renamed the demo dirs + manifest entries).

Frontend changes from QA agents were filtered: kept where they ported
LP-verbatim into ADK, replaced with direct LP pulls where the agent
had made ADK-specific adaptations (one Unicode-glyph case, one
stale-registry-slug case).

Not touched per blitz rules: shared_chat.py, registry.py, manifest.yaml,
src/app/api/copilotkit/route.ts.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-12 17:58:06 -07:00

3.8 KiB
Raw Permalink Blame History

QA: Sub-Agents — Google ADK

Prerequisites

  • Demo is deployed and accessible at /demos/subagents on the dashboard host
  • Agent backend is healthy (/api/health); GOOGLE_API_KEY is set (or GOOGLE_GEMINI_BASE_URL points at the aimock proxy)
  • A supervisor LlmAgent delegates to three sub-agents (research / writing / critique) via tools. Each delegation appends one completed entry to state["delegations"].

Test Steps

1. Basic Functionality

  • Navigate to /demos/subagents; verify the page renders within 3s with the "Sub-agent delegations" panel on the left and the CopilotChat on the right
  • Verify the delegation panel header reads "Sub-agent delegations" with a 0 calls counter on initial load
  • Verify the three sub-agent role indicator chips render at the top of the panel — Researcher, Writer, Critic — with dimmed (un-fired) styling
  • Verify the empty-state hint reads "Ask the supervisor to complete a task. Every sub-agent it calls will appear here."
  • Verify the chat input placeholder reads "Give the supervisor a task..."
  • Send "Hello" and verify the agent responds within 10s

2. Feature-Specific Checks

Suggestions

  • Verify all 3 suggestion pills are visible with verbatim titles:
    • "Write a blog post"
    • "Explain a topic"
    • "Summarize a topic"

Multi-Agent Delegation (Research → Writing → Critique)

  • Click "Write a blog post"
  • Within 60s verify the chat stream renders 3 inline activity cards in sequence:
    • data-testid="subagent-card-researcher" — Researcher activity card
    • data-testid="subagent-card-writer" — Writer activity card
    • data-testid="subagent-card-critic" — Critic activity card
  • Verify each card walks through statuses: starting → running → done (each card surfaces data-status="complete" once finished)
  • Verify each card's data-testid="subagent-result" contains real generated text (not "(empty)" and not the showcase intro boilerplate)
  • Verify the side-panel "Sub-agent delegations" list shows 3 delegation-entry cards (one per sub-agent), each labeled Research / Writing / Critique with the corresponding task
  • Verify the Supervisor running badge appears next to the header while the supervisor is active and disappears once it finishes

Active-Subagent Banner

  • During a delegation run, verify the sticky "active subagent" banner (data-testid="active-subagent-banner") appears at the top of the chat panel and names whichever sub-agent is currently running
  • Verify the banner disappears once the supervisor completes

Critic Loop Regression (one critic card per run)

  • Confirm that exactly 1 critic card renders per pill click — the count must be stable across a 5s dwell after completion
  • If a second critic card appears, the supervisor is re-entering the critic — file a bug (see subagents_agent.py instruction enforcing single-call-per-tool)

All Pills Run the Same Flow

  • Click "Explain a topic" — verify the same 3-card / 3-delegation pattern
  • Click "Summarize a topic" — verify the same 3-card / 3-delegation pattern

3. Error Handling

  • Send an empty message; verify it is a no-op
  • If a Gemini sub-agent call fails, verify the delegation log shows the error as a completed entry whose result contains the user-facing failure message (server logs preserve the traceback)
  • Verify no uncaught console errors during normal usage

Expected Results

  • Chat loads within 3 seconds; first delegation card appears within 60 seconds
  • Each of the 3 sub-agents fires exactly once per pill click
  • Card results are real generated text, not boilerplate
  • Side panel and inline chat stay synchronized via state["delegations"]
  • No console errors, no stuck-state, no critic-loop regression
Reference in New Issue View Git Blame Copy Permalink