Add examples#3
Conversation
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the You can disable this status message by setting the WalkthroughAdds three new example agents (Telegram, Streaming, LLM Judge), associated docs and configs. Introduces Telegram helper utilities and README. Updates Slack agent to use background processing via waitUntil. Extends agent registry and README entries. Minor README formatting tweaks, test-suite comment updates, env example note, and SDK version bump. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Telegram as Telegram
participant Agent as example-telegram
participant KV as KV Store
participant LLM as OpenAI
participant TGAPI as Telegram API
Telegram->>Agent: POST /webhook (update)
Note right of Agent: Validate + filter update
Agent-->>Telegram: 200 OK (immediate)
rect rgba(200,220,255,0.2)
Note over Agent: Background via waitUntil
Agent->>KV: get(chat_history)
KV-->>Agent: history[]
Agent->>KV: put(updated_history, TTL=86400)
Agent->>LLM: generate(reply | system+history+user)
LLM-->>Agent: text
Agent->>TGAPI: sendMessage(chat_id, text, reply_to?)
TGAPI-->>Agent: ok/json
Agent->>KV: put(history+assistant, TTL=86400)
end
sequenceDiagram
autonumber
actor Slack as Slack
participant Agent as example-slack
participant LLM as OpenAI
participant SlackAPI as Slack API
Slack->>Agent: Event/Webhook
Agent-->>Slack: 200 OK (verification/ack)
rect rgba(200,220,255,0.2)
Note over Agent: Background via waitUntil
Agent->>Agent: Build thread/history
Agent->>LLM: Generate reply
LLM-->>Agent: text
Agent->>SlackAPI: postMessage(thread_ts, text)
SlackAPI-->>Agent: ok
end
sequenceDiagram
autonumber
participant Client as Client
participant Agent as example-streaming
participant Other as example-chat
participant LLM as OpenAI
participant Streams as Stream Service
Client->>Agent: Prompt
alt "Agent Chain"
Agent->>Other: request(stream=true)
Other-->>Agent: stream
Agent-->>Client: resp.stream(stream)
else "LLM Streaming"
Agent->>Streams: create(llm-summary)
Agent-->>Client: JSON(stream info)
Agent->>LLM: streamText(...)
LLM-->>Agent: tokens
Agent->>Streams: pipe tokens
else "Manual Streaming"
Agent->>Streams: create(batch-processing, compressed)
Agent-->>Client: JSON(stream info)
loop items 1..100
Agent->>Streams: write(processed item)
end
Agent->>Streams: close()
else default
Agent-->>Client: Invalid input message
end
sequenceDiagram
autonumber
participant User as User
participant Judge as example-llm-judge
participant Provider as gateway-provider
participant LLM as OpenAI
User->>Judge: Request (prompt?)
Judge->>Provider: Generate Story A
Provider-->>Judge: text A
Judge->>Provider: Generate Story B
Provider-->>Judge: text B
Judge->>LLM: generateObject(schema: evaluation)
LLM-->>Judge: evaluation JSON
Judge-->>User: Markdown report
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Poem
Pre-merge checks and finishing touches❌ Failed checks (1 warning, 1 inconclusive)
✅ Passed checks (1 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
🧪 Preview Environments Are Live —
|
| IO Type | Access |
|---|---|
| ⏰ Cron | Coming Soon 🚀 |
🤖 io-api
| IO Type | Access |
|---|---|
| 🛜 API | POST https://agentuity.ai/api/agent_b61035c3341ac311663fad947c57eb7e/1dbcf43 |
📋 Curl Examples
ℹ️ Info: API & Webhooks requiring authentication will include placeholder tokens in the curl examples. Replace these with your actual credentials.
API Endpoint:
curl -X POST https://agentuity.ai/api/agent_b61035c3341ac311663fad947c57eb7e/1dbcf43 \
-H "Content-Type: application/json" \
-d '{}'
🤖 io-email
| IO Type | Access |
|---|---|
| Coming Soon 🚀 |
🤖 io-sms
| IO Type | Access |
|---|---|
| 📱 SMS | Coming Soon 🚀 |
🤖 io-webhook
| IO Type | Access |
|---|---|
| 🪝 Webhook | POST https://agentuity.ai/webhook/agent_b45feaf99bbdb042b3e8838c8544abe7/1dbcf43 |
📋 Curl Examples
ℹ️ Info: API & Webhooks requiring authentication will include placeholder tokens in the curl examples. Replace these with your actual credentials.
Webhook Endpoint:
curl -X POST https://agentuity.ai/webhook/agent_b45feaf99bbdb042b3e8838c8544abe7/1dbcf43 \
-H "Content-Type: application/json" \
-d '{}'
🤖 observability-logging
No input channels configured — ➕ Add I/O
🤖 observability-tracing
No input channels configured — ➕ Add I/O
🤖 handler-request
No input channels configured — ➕ Add I/O
🤖 handler-response
| IO Type | Access |
|---|---|
| 🛜 API | POST https://agentuity.ai/api/agent_a2d37fcdd47a1df58c621022909932ba/1dbcf43 |
📋 Curl Examples
ℹ️ Info: API & Webhooks requiring authentication will include placeholder tokens in the curl examples. Replace these with your actual credentials.
API Endpoint:
curl -X POST https://agentuity.ai/api/agent_a2d37fcdd47a1df58c621022909932ba/1dbcf43 \
-H "Content-Type: application/json" \
-d '{}'
🤖 handler-context
No input channels configured — ➕ Add I/O
🤖 storage-vector
No input channels configured — ➕ Add I/O
🤖 example-discord
No input channels configured — ➕ Add I/O
🤖 example-chat
No input channels configured — ➕ Add I/O
🤖 example-slack
| IO Type | Access |
|---|---|
| 🛜 API | POST https://agentuity.ai/api/agent_2ddc01490cd360ec3d1d594999bda18e/1dbcf43 |
📋 Curl Examples
ℹ️ Info: API & Webhooks requiring authentication will include placeholder tokens in the curl examples. Replace these with your actual credentials.
API Endpoint:
curl -X POST https://agentuity.ai/api/agent_2ddc01490cd360ec3d1d594999bda18e/1dbcf43 \
-H "Content-Type: application/json" \
-d '{}'
🤖 example-composio
No input channels configured — ➕ Add I/O
🤖 example-telegram
| IO Type | Access |
|---|---|
| 🪝 Webhook | POST https://agentuity.ai/webhook/agent_52f73139a881ee0e1cdfafb3c6404e70/1dbcf43 |
📋 Curl Examples
ℹ️ Info: API & Webhooks requiring authentication will include placeholder tokens in the curl examples. Replace these with your actual credentials.
Webhook Endpoint:
curl -X POST https://agentuity.ai/webhook/agent_52f73139a881ee0e1cdfafb3c6404e70/1dbcf43 \
-H "Content-Type: application/json" \
-d '{}'📊 Resource Allocation
| Resource | Allocation |
|---|---|
| 🧠 Memory | 1,000 B |
| 🧮 CPU | 0.5 m |
| 💾 Disk | 150 kB |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (4)
.github/example-telegram/telegram-setup-1-api-endpoint.pngis excluded by!**/*.png.github/example-telegram/telegram-setup-2-bot-father.pngis excluded by!**/*.png.github/example-telegram/telegram-setup-3-token.pngis excluded by!**/*.pngbun.lockis excluded by!**/*.lock
📒 Files selected for processing (7)
.env.example(1 hunks)agentuity.yaml(1 hunks)package.json(1 hunks)src/agents/example-telegram/README.md(1 hunks)src/agents/example-telegram/index.ts(1 hunks)src/agents/example-telegram/telegram.ts(1 hunks)src/agents/test-suite/index.ts(2 hunks)
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Prefer loading types from the@agentuity/sdkpackage (from node_modules)
Import types from@agentuity/sdk
Use structured error handling with try/catch blocks
Files:
src/agents/example-telegram/telegram.tssrc/agents/example-telegram/index.tssrc/agents/test-suite/index.ts
🧠 Learnings (2)
📚 Learning: 2025-09-03T20:25:54.918Z
Learnt from: CR
PR: agentuity/kitchen-sink-ts#0
File: AGENTS.md:0-0
Timestamp: 2025-09-03T20:25:54.918Z
Learning: Applies to **/*.{ts,tsx} : Import types from `agentuity/sdk`
Applied to files:
package.json
📚 Learning: 2025-09-03T20:25:54.918Z
Learnt from: CR
PR: agentuity/kitchen-sink-ts#0
File: AGENTS.md:0-0
Timestamp: 2025-09-03T20:25:54.918Z
Learning: Applies to **/*.{ts,tsx} : Prefer loading types from the `agentuity/sdk` package (from node_modules)
Applied to files:
package.json
🔇 Additional comments (5)
package.json (1)
31-31: Verify @agentuity/sdk 0.0.146 before merging.
^0.0.146wasn’t published the last time I checked, so installs could fail if the tag still isn’t live. Please confirm the package exists (or bump to the latest released version).Based on learnings
.env.example (1)
14-15: Nice touch documenting the Telegram token.The placeholder keeps the sample environment file aligned with the new bot example so folks can configure it without guesswork.
agentuity.yaml (1)
148-153: New agent entries are wired in cleanly.
example-teamsandexample-telegramnow surface like the other kitchen sink agents, so they’ll be available to the CLI and dashboard without extra manual edits.src/agents/example-telegram/README.md (1)
1-117: README comprehensively guides the Telegram setup.End-to-end steps (deploy, BotFather, env vars, webhook curl, testing, limits) cover every piece needed to bring the bot online—great developer experience.
src/agents/test-suite/index.ts (1)
68-110: Appreciate the inline rationale for skipping Telegram in automation.Leaving the entries commented with context about webhook constraints prevents accidental load while still documenting how to exercise the agent.
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (1)
src/agents/example-telegram/README.md (1)
35-37: Add a language hint to the.envfenced block.markdownlint is flagging the unlabeled fence. Tag it as
dotenv(or another appropriate lexer) so the lint check passes and readers get syntax highlighting.-``` +```dotenv TELEGRAM_BOT_TOKEN=1234567890:ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghi</blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used**: CodeRabbit UI **Review profile**: CHILL **Plan**: Pro <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 7ba7ea1a339b3142b75f01e23e79bbe4987b7090 and 1dbcf43feb4ece37cd82da7b7952ebd453faa1a4. </details> <details> <summary>📒 Files selected for processing (4)</summary> * `agentuity.yaml` (1 hunks) * `src/agents/example-telegram/README.md` (1 hunks) * `src/agents/example-telegram/index.ts` (1 hunks) * `src/agents/example-telegram/telegram.ts` (1 hunks) </details> <details> <summary>🚧 Files skipped from review as they are similar to previous changes (1)</summary> * src/agents/example-telegram/telegram.ts </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>📓 Path-based instructions (1)</summary> <details> <summary>**/*.{ts,tsx}</summary> **📄 CodeRabbit inference engine (AGENTS.md)** > `**/*.{ts,tsx}`: Prefer loading types from the `@agentuity/sdk` package (from node_modules) > Import types from `@agentuity/sdk` > Use structured error handling with try/catch blocks Files: - `src/agents/example-telegram/index.ts` </details> </details><details> <summary>🧬 Code graph analysis (1)</summary> <details> <summary>src/agents/example-telegram/index.ts (3)</summary><blockquote> <details> <summary>src/agents/example-slack/index.ts (1)</summary> * `Agent` (23-119) </details> <details> <summary>src/agents/example-chat/index.ts (1)</summary> * `Agent` (11-83) </details> <details> <summary>src/agents/example-telegram/telegram.ts (2)</summary> * `TelegramUpdate` (4-7) * `sendTelegramMessage` (28-57) </details> </blockquote></details> </details><details> <summary>🪛 markdownlint-cli2 (0.18.1)</summary> <details> <summary>src/agents/example-telegram/README.md</summary> 35-35: Fenced code blocks should have a language specified (MD040, fenced-code-language) </details> </details> </details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
- Remove Teams (backlog until SDK changes)
parteeksingh24
force-pushed
the
add-examples
branch
from
1dbcf43 to
3fa9142
Compare
- Use `waitUntil` for background processing - Add bot profile null safety
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
src/agents/example-telegram/README.md (1)
35-37: Add language specifier to fenced code block.The code block is missing a language identifier. Adding
bashorenvwill enable syntax highlighting and improve readability.Apply this diff:
For local development, add to your `.env` file: -``` +```env TELEGRAM_BOT_TOKEN=1234567890:ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghiAs per static analysis hints. </blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used**: CodeRabbit UI **Review profile**: CHILL **Plan**: Pro <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 1dbcf43feb4ece37cd82da7b7952ebd453faa1a4 and a97aa26e90869e5f24c36bd78f7788b3172e55c8. </details> <details> <summary>⛔ Files ignored due to path filters (4)</summary> * `.github/example-telegram/telegram-setup-1-api-endpoint.png` is excluded by `!**/*.png` * `.github/example-telegram/telegram-setup-2-bot-father.png` is excluded by `!**/*.png` * `.github/example-telegram/telegram-setup-3-token.png` is excluded by `!**/*.png` * `bun.lock` is excluded by `!**/*.lock` </details> <details> <summary>📒 Files selected for processing (13)</summary> * `.env.example` (1 hunks) * `README.md` (1 hunks) * `agentuity.yaml` (1 hunks) * `src/agents/example-llm-judge/index.ts` (1 hunks) * `src/agents/example-llm-judge/story-eval.ts` (1 hunks) * `src/agents/example-slack/index.ts` (2 hunks) * `src/agents/example-slack/slack.ts` (1 hunks) * `src/agents/example-streaming/index.ts` (1 hunks) * `src/agents/example-streaming/sample-data.json` (1 hunks) * `src/agents/example-telegram/README.md` (1 hunks) * `src/agents/example-telegram/index.ts` (1 hunks) * `src/agents/example-telegram/telegram.ts` (1 hunks) * `src/agents/test-suite/index.ts` (2 hunks) </details> <details> <summary>✅ Files skipped from review due to trivial changes (2)</summary> * src/agents/example-streaming/sample-data.json * README.md </details> <details> <summary>🚧 Files skipped from review as they are similar to previous changes (2)</summary> * src/agents/example-telegram/telegram.ts * src/agents/test-suite/index.ts </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>📓 Path-based instructions (1)</summary> <details> <summary>**/*.{ts,tsx}</summary> **📄 CodeRabbit inference engine (AGENTS.md)** > `**/*.{ts,tsx}`: Prefer loading types from the `@agentuity/sdk` package (from node_modules) > Import types from `@agentuity/sdk` > Use structured error handling with try/catch blocks Files: - `src/agents/example-llm-judge/story-eval.ts` - `src/agents/example-llm-judge/index.ts` - `src/agents/example-slack/slack.ts` - `src/agents/example-slack/index.ts` - `src/agents/example-streaming/index.ts` - `src/agents/example-telegram/index.ts` </details> </details><details> <summary>🧬 Code graph analysis (4)</summary> <details> <summary>src/agents/example-llm-judge/index.ts (2)</summary><blockquote> <details> <summary>src/agents/test-suite/index.ts (2)</summary> * `Agent` (34-142) * `welcome` (22-26) </details> <details> <summary>src/agents/example-llm-judge/story-eval.ts (2)</summary> * `evaluationSchema` (16-20) * `formatReport` (25-56) </details> </blockquote></details> <details> <summary>src/agents/example-slack/index.ts (2)</summary><blockquote> <details> <summary>src/agents/example-slack/slack.ts (2)</summary> * `verifySlackWebhook` (22-121) * `SlackAgentRequest` (11-19) </details> <details> <summary>src/lib/utils.ts (1)</summary> * `handleError` (105-114) </details> </blockquote></details> <details> <summary>src/agents/example-streaming/index.ts (1)</summary><blockquote> <details> <summary>src/lib/utils.ts (1)</summary> * `handleHelpMessage` (17-103) </details> </blockquote></details> <details> <summary>src/agents/example-telegram/index.ts (2)</summary><blockquote> <details> <summary>src/agents/example-chat/index.ts (1)</summary> * `Agent` (11-83) </details> <details> <summary>src/agents/example-telegram/telegram.ts (2)</summary> * `TelegramUpdate` (4-7) * `sendTelegramMessage` (28-57) </details> </blockquote></details> </details><details> <summary>🪛 markdownlint-cli2 (0.18.1)</summary> <details> <summary>src/agents/example-telegram/README.md</summary> 35-35: Fenced code blocks should have a language specified (MD040, fenced-code-language) </details> </details> </details> <details> <summary>🔇 Additional comments (26)</summary><blockquote> <details> <summary>src/agents/example-slack/slack.ts (1)</summary><blockquote> `96-96`: **LGTM! Clean refactor to optional chaining.** The change simplifies the bot profile check while preserving the same guard logic against infinite loops. </blockquote></details> <details> <summary>src/agents/example-slack/index.ts (5)</summary><blockquote> `29-32`: **LGTM! Proper manual trigger guard.** This prevents unintended manual invocations of a webhook-only agent. --- `34-43`: **LGTM! Correct webhook verification.** The verification call properly handles Slack challenges, signature validation, and bot message filtering before processing. --- `45-52`: **LGTM! Proper non-blocking webhook pattern.** The early `200 OK` response with `ctx.waitUntil` for background processing aligns with Slack's webhook requirements and matches the pattern used in other example agents (e.g., example-telegram). --- `97-101`: **LGTM! Proper background task error containment.** Errors inside the background task are logged and handled gracefully without throwing to the outer scope, which is the correct pattern for `ctx.waitUntil` tasks. --- `86-86`: **No action needed: ‘gpt-5-mini’ is a supported OpenAI model identifier.** </blockquote></details> <details> <summary>src/agents/example-streaming/index.ts (6)</summary><blockquote> `1-20`: **LGTM! Clean setup and boilerplate.** Imports follow coding guidelines (types from `@agentuity/sdk`), and the help message handler is properly integrated. --- `29-105`: **LGTM! Solid manual streaming pattern with proper resource cleanup.** The manual progress example demonstrates correct usage of `ctx.stream.create()`, `WritableStream` writer lifecycle (`getWriter()`, `releaseLock()`), and background processing with `ctx.waitUntil()`. The `finally` block ensures the stream is always closed. --- `128-146`: **LGTM! Correct background streaming with pipeTo.** The pattern properly uses `ctx.waitUntil()` for background LLM streaming, pipes the result to the created stream, and ensures the stream is closed even on error. --- `170-203`: **LGTM! Clean agent-to-agent streaming.** The agent chain example correctly uses `ctx.getAgent()`, calls another agent with `run()`, retrieves the response stream, and forwards it with `resp.stream()`. Error handling is appropriate. --- `208-226`: **LGTM! Well-structured welcome metadata.** The welcome message provides clear context and links to the streaming guide, with well-defined prompts for each streaming pattern. --- `108-167`: **‘gpt-5-nano’ model identifier is correct.** Confirmed that OpenAI’s GPT-5 family includes this variant. </blockquote></details> <details> <summary>.env.example (1)</summary><blockquote> `14-16`: **LGTM!** The Telegram bot token placeholder follows the established pattern for example agent environment variables and is properly commented out. </blockquote></details> <details> <summary>src/agents/example-llm-judge/story-eval.ts (1)</summary><blockquote> `1-56`: **LGTM!** The evaluation schema is well-structured with proper Zod validation constraints (min/max for scoring), clear field descriptions, and a clean formatting function. The schema design appropriately separates story evaluation from verdict, making it easy to consume in the LLM-as-a-judge workflow. </blockquote></details> <details> <summary>agentuity.yaml (1)</summary><blockquote> `148-156`: **LGTM!** The three new agent registrations follow the established format and align with the new agent implementations added in this PR. The descriptions are clear and consistent with the existing agent entries. </blockquote></details> <details> <summary>src/agents/example-llm-judge/index.ts (2)</summary><blockquote> `1-10`: **LGTM!** The imports and function signature properly use types from `@agentuity/sdk` and follow the coding guidelines. --- `72-90`: **LGTM!** The welcome function provides clear documentation and useful example prompts for testing the agent. </blockquote></details> <details> <summary>src/agents/example-telegram/index.ts (9)</summary><blockquote> `1-4`: **LGTM!** The imports are well-organized and follow the coding guidelines by loading types from `@agentuity/sdk`. All imported modules are used appropriately in the agent implementation. --- `11-14`: **LGTM!** The manual trigger rejection is clear and appropriate for a webhook-only agent. --- `16-22`: **LGTM!** The webhook parsing and message filtering logic correctly handles non-text and bot messages, preventing the agent from processing irrelevant updates or creating feedback loops. --- `24-29`: **LGTM!** The fire-and-forget pattern is appropriate for Telegram webhook handlers, ensuring a fast 200 OK response while processing the message asynchronously in the background. --- `31-66`: **LGTM!** The conversation history management is well-implemented: - Unique per-chat storage with appropriate error handling - Reasonable 10-message limit to control context size - 24-hour TTL aligns with typical chat session lifetimes - The non-null assertion on line 56 is safe given the earlier guard --- `76-89`: **LGTM!** The token validation correctly addresses the previous review comment by failing fast with a clear error message when `TELEGRAM_BOT_TOKEN` is missing. The validated token is then safely passed to `sendTelegramMessage` with proper threading via `replyToMessageId`. --- `91-103`: **LGTM!** The assistant response is correctly appended to history with a matching Unix timestamp format, and the error handling follows the coding guidelines by using structured try/catch with proper logging via `ctx.logger`. --- `105-108`: **LGTM!** The welcome function provides clear, informative metadata about the Telegram agent with appropriate styling and helpful notes about testing limitations in DevMode. --- `68-74`: **GPT-5 mini is a valid model name** The `gpt-5-mini` variant is officially supported in the OpenAI API alongside `gpt-5` and `gpt-5-nano` ([openai.com](https://openai.com/te-IN/index/introducing-gpt-5-for-developers/?utm_source=openai)). No changes required. > Likely an incorrect or invalid review comment. </blockquote></details> </blockquote></details> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
| try { | ||
| // Get the prompt from request, or use a default | ||
| const prompt = | ||
| (await req.data.text()) || | ||
| 'Write a short story about an AI learning to paint'; | ||
|
|
||
| ctx.logger.info('Starting LLM-as-a-judge evaluation'); | ||
|
|
||
| // Get stories from gateway-provider | ||
| const gatewayAgent = await ctx.getAgent({ name: 'gateway-provider' }); | ||
| const stories = await gatewayAgent.run({ data: prompt }); | ||
| const storiesText = await stories.data.text(); | ||
|
|
||
| ctx.logger.debug('Received stories from gateway-provider'); | ||
|
|
||
| // Create evaluation prompt | ||
| const evaluationPrompt = ` | ||
| You are evaluating two AI-generated short stories. | ||
|
|
||
| Here are the stories: | ||
|
|
||
| ${storiesText} | ||
|
|
||
| Extract each story text: | ||
| - OpenAI story: appears after "### OpenAI (GPT-5 Nano)" | ||
| - Google story: appears after "### Google (Gemini 2.0 Flash)" | ||
|
|
||
| For each story, provide: | ||
| 1. Creativity score (1-10): How original and imaginative is it? | ||
| 2. Quality score (1-10): Overall writing quality | ||
| 3. Strengths: What works well (1-2 sentences) | ||
|
|
||
| Finally, provide a verdict declaring which story is better and why (2-3 sentences).`; | ||
|
|
||
| ctx.logger.info('Generating structured evaluation'); | ||
|
|
||
| // Generate structured evaluation | ||
| const evaluation = await generateObject({ | ||
| model: openai('gpt-5-nano'), | ||
| schema: evaluationSchema, | ||
| system: | ||
| 'You are a literary critic evaluating short AI-generated stories.', | ||
| prompt: evaluationPrompt, | ||
| }); | ||
|
|
||
| // Log key metrics | ||
| ctx.logger.debug('Evaluation scores', { | ||
| openai: evaluation.object.openai.quality, | ||
| google: evaluation.object.google.quality, | ||
| }); | ||
|
|
||
| // Return formatted report | ||
| return resp.markdown(formatReport(evaluation.object)); | ||
| } catch (error) { | ||
| ctx.logger.error('Error in LLM judge evaluation:', error); | ||
| return resp.text( | ||
| 'Sorry, there was an error running the evaluation. Please ensure the gateway-provider agent is available.' | ||
| ); | ||
| } | ||
| } |
There was a problem hiding this comment.
Verify the model name and consider more robust parsing.
The implementation follows structured error handling guidelines and correctly integrates with the evaluation schema. However, there are two concerns:
-
Model name verification needed: Line 49 uses
openai('gpt-5-nano'), which doesn't match known OpenAI model naming conventions (typicallygpt-4,gpt-4-turbo,gpt-3.5-turbo, etc.). If this model doesn't exist, the agent will fail at runtime. -
Fragile format coupling: Lines 35-36 hardcode markdown section headers to extract stories from
gateway-provideroutput. If the upstream agent changes its output format, the evaluation prompt won't correctly identify the stories, leading to poor extraction results or evaluation failures.
For issue 1, verify the correct model name:
What OpenAI models are available through the OpenAI API as of 2025? Is gpt-5-nano a valid model?
For issue 2, consider adding format validation or using a more robust parsing approach:
// Get stories from gateway-provider
const gatewayAgent = await ctx.getAgent({ name: 'gateway-provider' });
const stories = await gatewayAgent.run({ data: prompt });
const storiesText = await stories.data.text();
ctx.logger.debug('Received stories from gateway-provider');
+
+ // Validate expected format
+ if (!storiesText.includes('### OpenAI') || !storiesText.includes('### Google')) {
+ throw new Error('Unexpected format from gateway-provider agent');
+ }
// Create evaluation prompt🤖 Prompt for AI Agents
In src/agents/example-llm-judge/index.ts around lines 11 to 70, the review flags
two problems: the call to openai('gpt-5-nano') may reference a non-existent
model and the evaluation prompt relies on brittle hard-coded markdown headers to
locate stories. Fix by replacing the model name with a verified supported model
(e.g., gpt-4 or gpt-4o or configurable via env/ctx config) or make the model
string configurable and validate it at startup; and make story extraction
resilient by having the gateway-provider return structured JSON (preferred) or
include explicit delimiters you parse robustly (e.g., regex with safe fallbacks
and validation checks) and add error handling/logging when parsing fails so the
evaluation uses validated story texts instead of assuming exact headers.
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (2)
src/agents/example-telegram/index.ts (1)
16-22: Consider error handling for webhook parsing.The webhook parsing at line 17 may throw if the payload is malformed. While Telegram webhooks should be reliable, adding a try-catch around the parsing and filtering logic would prevent unhandled exceptions from reaching the platform.
Apply this diff to add error handling:
+ try { // Parse webhook update const update = await req.data.object<TelegramUpdate>(); // Filter non-text messages and bot messages if (!update?.message?.text || update.message.from?.is_bot) { return resp.text('OK'); } + } catch (error) { + ctx.logger.error('Error parsing Telegram webhook:', error); + return resp.text('OK'); + }src/agents/example-telegram/README.md (1)
108-111: Rate limits confirmed. The documented limits (1 message/sec per chat, 20 messages/min per group, and ~30 messages/sec overall for free-tier broadcasts) match the current Telegram Bot API. Optionally, you may note that enabling Paid Broadcasts raises this up to 1,000 messages/sec. citeturn0fetch0turn0fetch0
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (4)
src/agents/example-composio/README.md(1 hunks)src/agents/example-discord/README.md(1 hunks)src/agents/example-telegram/README.md(1 hunks)src/agents/example-telegram/index.ts(1 hunks)
✅ Files skipped from review due to trivial changes (1)
- src/agents/example-discord/README.md
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Prefer loading types from the@agentuity/sdkpackage (from node_modules)
Import types from@agentuity/sdk
Use structured error handling with try/catch blocks
Files:
src/agents/example-telegram/index.ts
🧬 Code graph analysis (1)
src/agents/example-telegram/index.ts (2)
src/agents/example-chat/index.ts (1)
Agent(11-83)src/agents/example-telegram/telegram.ts (2)
TelegramUpdate(4-7)sendTelegramMessage(28-57)
🔇 Additional comments (6)
src/agents/example-composio/README.md (1)
44-46: LGTM! Documentation formatting improvement.The switch to an
env-tagged code block improves syntax highlighting and aligns with the documentation standards being adopted across other agent READMEs in this PR.src/agents/example-telegram/README.md (2)
1-66: LGTM! Clear and comprehensive setup guide.The step-by-step instructions are well-structured and cover all necessary configuration points including security best practices. The inclusion of both local development and production deployment options is helpful.
73-88: LGTM! Accurate functional description.The description correctly reflects the agent's capabilities and aligns with the implementation in
index.ts, including the 10-message history limit and supported chat types.src/agents/example-telegram/index.ts (3)
6-14: LGTM! Proper type usage and clear trigger handling.The function signature correctly uses types from
@agentuity/sdkas per coding guidelines, and the manual trigger guard provides clear feedback for webhook-only operation.
34-66: LGTM! Solid KV history management with proper error handling.The conversation history implementation follows established patterns from
example-chat, includes structured error handling per coding guidelines, and correctly manages message history with appropriate TTL. The inline type definition is acceptable for this single-use case.
105-108: LGTM! Appropriate welcome message for platform-specific agent.The welcome export correctly explains the agent's purpose and limitations regarding DevMode testing, which is appropriate for a webhook-based Telegram bot.
There was a problem hiding this comment.
Actionable comments posted: 2
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
src/agents/example-streaming/index.ts(1 hunks)src/agents/example-telegram/index.ts(1 hunks)
🚧 Files skipped from review as they are similar to previous changes (1)
- src/agents/example-telegram/index.ts
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Prefer loading types from the@agentuity/sdkpackage (from node_modules)
Import types from@agentuity/sdk
Use structured error handling with try/catch blocks
Files:
src/agents/example-streaming/index.ts
🧬 Code graph analysis (1)
src/agents/example-streaming/index.ts (1)
src/lib/utils.ts (1)
handleHelpMessage(17-103)
🔇 Additional comments (5)
src/agents/example-streaming/index.ts (5)
1-5: LGTM! Clean imports following project guidelines.Types are correctly imported from
@agentuity/sdkas per coding guidelines, and AI SDK usage aligns with best practices from the learnings.
7-20: LGTM! Proper agent structure with help handling.Function signature is correct and the early help message handling pattern is clean and well-implemented.
28-106: Well-structured manual streaming example with proper resource cleanup.The pattern correctly demonstrates
ctx.stream.create()with manual writes, proper use ofctx.waitUntil()for background processing, and excellent resource management withwriter.releaseLock()andstream.close()in the finally block. The 4-second delay at line 83 is appropriate for this example code to demonstrate streaming behavior.
170-204: Excellent agent-to-agent streaming implementation.The agent chain pattern is well-implemented with proper error handling, informative logging, and correct usage of
ctx.getAgent(), agentrun(), andresp.stream()for forwarding. The error message helpfully mentions the dependency onexample-chatagent.
209-227: Well-crafted welcome function with clear UX.The welcome message provides helpful context with styled formatting and documentation links. The three prompts correctly correspond to the implemented streaming patterns with appropriate content types.
There was a problem hiding this comment.
Actionable comments posted: 2
📜 Review details
Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro
⛔ Files ignored due to path filters (1)
bun.lockis excluded by!**/*.lock
📒 Files selected for processing (2)
package.json(2 hunks)src/agents/example-streaming/index.ts(1 hunks)
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{ts,tsx}
📄 CodeRabbit inference engine (AGENTS.md)
**/*.{ts,tsx}: Prefer loading types from the@agentuity/sdkpackage (from node_modules)
Import types from@agentuity/sdk
Use structured error handling with try/catch blocks
Files:
src/agents/example-streaming/index.ts
🧠 Learnings (2)
📚 Learning: 2025-09-03T20:25:54.918Z
Learnt from: CR
PR: agentuity/kitchen-sink-ts#0
File: AGENTS.md:0-0
Timestamp: 2025-09-03T20:25:54.918Z
Learning: Applies to **/*.{ts,tsx} : Import types from `agentuity/sdk`
Applied to files:
package.json
📚 Learning: 2025-09-03T20:25:54.918Z
Learnt from: CR
PR: agentuity/kitchen-sink-ts#0
File: AGENTS.md:0-0
Timestamp: 2025-09-03T20:25:54.918Z
Learning: Applies to **/*.{ts,tsx} : Prefer loading types from the `agentuity/sdk` package (from node_modules)
Applied to files:
package.json
🧬 Code graph analysis (1)
src/agents/example-streaming/index.ts (1)
src/lib/utils.ts (1)
handleHelpMessage(17-103)
| }, | ||
| "dependencies": { | ||
| "@agentuity/sdk": "^0.0.146", | ||
| "@agentuity/sdk": "^0.0.147", |
There was a problem hiding this comment.
Dependency version 0.0.147 is not available on npm
npm still lists 0.0.144 as the latest publish for @agentuity/sdk, so installing ^0.0.147 will currently fail. Please either wait for 0.0.147 to be published or pin to the latest available release.
(npmjs.com)
🤖 Prompt for AI Agents
In package.json around line 31, the dependency "@agentuity/sdk": "^0.0.147"
references a version not published to npm; update the version to a published
release (e.g., change to "^0.0.144" or pin to "0.0.144") or revert to the
previous working version, then run npm install (or yarn) to verify resolution
and update lockfile accordingly.
- Add Checkly (heartbeat) monitoring for out-of-band agents - Improve Slack bot (`app_mention` events, retry handling, Groq) - Update streaming example (list/delete methods) - Add example images for Teams manifest (`color.png`, `outline.png`)
1 participant
Summary by CodeRabbit