feat(skills): Surface Eve, Flue, and Cloudflare AI - #341
Conversation
Add detection and setup guidance for Eve, Flue, the Cloudflare Agents SDK, and Workers AI to the shared instrumentation skill used by every plugin build. Refs TET-2938 Co-Authored-By: OpenAI GPT-5.6 Codex <noreply@openai.com>
Route Cloudflare Agents setup to the existing Durable Objects reference, clarify conversation ID ownership, and detect the current Vercel AI and Google GenAI packages. Refs TET-2938 Co-Authored-By: OpenAI GPT-5.6 Codex <noreply@openai.com>
Base Eve and Flue routing on detected project state, and model Cloudflare host, model, and conversation instrumentation as composable layers. Refs TET-2938 Co-Authored-By: OpenAI GPT-5.6 Codex <noreply@openai.com>
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
Run framework ownership detection for every instrumentation scope and keep Eve's OTLP exporter mutually exclusive with the Node SDK in the same agent runtime. Refs TET-2938 Co-Authored-By: OpenAI GPT-5.6 Codex <noreply@openai.com>
Route fresh Eve OTLP setup through project provisioning, the framework installer, configuration, and AI-span verification before skipping the generic SDK flow. Refs TET-2938 Co-Authored-By: OpenAI GPT-5.6 Codex <noreply@openai.com>
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 25a4b0a. Configure here.
| runtime. For a fresh setup, run Step 2 of `first-error-setup.md` to select or create | ||
| the Sentry project. Then follow the detected platform’s `ai-monitoring.md` Eve | ||
| section through `eve add instrumentation/sentry`, configuration, and AI-span | ||
| verification. Do not continue to the generic SDK install or error verification; this |
There was a problem hiding this comment.
Eve setup uses the wrong platform
Medium Severity
The Eve OTLP route follows the detected platform’s ai-monitoring.md Eve section, but that section exists only under the Node SDK. Eve plus Next.js is an official layout and resolves to nextjs, so the route cannot find Eve setup and can apply generic Next.js AI instrumentation instead.
Reviewed by Cursor Bugbot for commit 25a4b0a. Configure here.
| Never create a second Sentry initialization, OTLP exporter, or AI span producer. | ||
|
|
||
| For **add a signal**, after completing any framework-owned handoff above, preserve the | ||
| selected base install and go to Step 3 for the requested signal. |
There was a problem hiding this comment.
Add-a-signal repeats completed Eve setup
Medium Severity
Add-a-signal now finishes the Eve OTLP handoff, which already runs eve add instrumentation/sentry and AI-span verification, then still continues to Step 3 to wire the same signal. That second pass can add generic Node provider integrations on top of Eve’s exporter.
Additional Locations (1)
Reviewed by Cursor Bugbot for commit 25a4b0a. Configure here.
2 participants
The shared
sentry-instrumentskill now detects and configures Eve, Flue, the Cloudflare Agents SDK, and Workers AI. Eve and Flue route through their framework-owned setup commands to avoid duplicate spans, while the Cloudflare guidance covers automatic Workers AI instrumentation, agent class wrapping, conversation IDs, privacy controls, verification, and common failures.Every plugin build hydrates this shared skill and its Node.js and Cloudflare references, so the new integrations are available across the Agent Plugins, Claude, Codex, Cursor, and Grok distributions without plugin-specific manifests.
closes TET-2938 (Linear issue).