feat(github): add issue templates for automatic triage

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -0,0 +1,117 @@
+name: Bug report
+description: Report broken behavior, regressions, crashes, or reliability issues.
+title: "[Bug]: "
+labels:
+  - bug
+  - needs-triage
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Use this form for broken behavior, regressions, crashes, or reliability problems.
+        Search existing issues first and keep the report focused on one problem.
+
+  - type: checkboxes
+    id: checks
+    attributes:
+      label: Before submitting
+      options:
+        - label: I searched existing issues and did not find a duplicate.
+          required: true
+        - label: I included enough detail to reproduce or investigate the problem.
+          required: true
+
+  - type: dropdown
+    id: area
+    attributes:
+      label: Area
+      description: Which part of the project is affected?
+      options:
+        - apps/web
+        - apps/server
+        - apps/desktop
+        - packages/contracts or packages/shared
+        - Build, CI, or release tooling
+        - Docs
+        - Not sure
+    validations:
+      required: true
+
+  - type: textarea
+    id: summary
+    attributes:
+      label: Summary
+      description: What broke? Include the first visible symptom.
+      placeholder: Sending a follow-up prompt after reconnecting leaves the thread stuck in "starting".
+    validations:
+      required: true
+
+  - type: textarea
+    id: steps
+    attributes:
+      label: Steps to reproduce
+      description: Provide a minimal, deterministic repro.
+      placeholder: |
+        1. Start the app with ...
+        2. Open ...
+        3. Disconnect the network
+        4. Reconnect and send another prompt
+    validations:
+      required: true
+
+  - type: textarea
+    id: expected
+    attributes:
+      label: Expected behavior
+      placeholder: The thread should resume normally and accept the new message.
+    validations:
+      required: true
+
+  - type: textarea
+    id: actual
+    attributes:
+      label: Actual behavior
+      placeholder: The composer clears, no new turn starts, and the session stays in a pending state.
+    validations:
+      required: true
+
+  - type: dropdown
+    id: impact
+    attributes:
+      label: Impact
+      description: How bad is this in practice?
+      options:
+        - Blocks work completely
+        - Major degradation or frequent failure
+        - Minor bug or occasional failure
+        - Cosmetic issue
+    validations:
+      required: true
+
+  - type: input
+    id: version
+    attributes:
+      label: Version or commit
+      description: Release version, commit SHA, or branch name if known.
+      placeholder: main @ abc1234
+
+  - type: input
+    id: environment
+    attributes:
+      label: Environment
+      description: OS, browser or desktop app version, Node/Bun version, provider/model if relevant.
+      placeholder: macOS 15.3, Chrome 135, Bun 1.3.9, Codex app-server ...
+
+  - type: textarea
+    id: logs
+    attributes:
+      label: Logs, stack traces, or screenshots
+      description: Paste the most relevant output only. Redact secrets.
+      render: shell
+
+  - type: textarea
+    id: workaround
+    attributes:
+      label: Workaround
+      description: If you found a temporary workaround, include it.
+      placeholder: Restarting the provider session clears the stuck state.

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -0,0 +1,101 @@
+name: Feature request
+description: Propose a scoped improvement or new capability.
+title: "[Feature]: "
+labels:
+  - enhancement
+  - needs-triage
+body:
+  - type: markdown
+    attributes:
+      value: |
+        Use this form for new capabilities or meaningful improvements to existing behavior.
+        This repo is still early. Small, concrete requests that clearly explain the problem and scope are much easier to evaluate.
+
+  - type: checkboxes
+    id: checks
+    attributes:
+      label: Before submitting
+      options:
+        - label: I searched existing issues and did not find a duplicate.
+          required: true
+        - label: I am describing a concrete problem or use case, not just a vague idea.
+          required: true
+
+  - type: dropdown
+    id: area
+    attributes:
+      label: Area
+      description: Which part of the project would this change affect?
+      options:
+        - apps/web
+        - apps/server
+        - apps/desktop
+        - packages/contracts or packages/shared
+        - Build, CI, or release tooling
+        - Docs
+        - Not sure
+    validations:
+      required: true
+
+  - type: textarea
+    id: problem
+    attributes:
+      label: Problem or use case
+      description: What are you trying to do? What is hard, slow, or impossible today?
+      placeholder: I want to reconnect to an existing provider session after a browser refresh without losing the current thread state.
+    validations:
+      required: true
+
+  - type: textarea
+    id: proposal
+    attributes:
+      label: Proposed solution
+      description: Describe the behavior, API, or UX you want.
+      placeholder: Persist enough session metadata so the client can discover and reattach to the active provider session on load.
+    validations:
+      required: true
+
+  - type: textarea
+    id: value
+    attributes:
+      label: Why this matters
+      description: Who benefits, and what outcome does this unlock?
+      placeholder: This would make reconnects predictable during network drops and reduce accidental duplicate sessions.
+    validations:
+      required: true
+
+  - type: textarea
+    id: scope
+    attributes:
+      label: Smallest useful scope
+      description: What is the narrowest version of this request that would still solve your problem?
+      placeholder: A first pass only needs to support restoring the active session for the current thread.
+    validations:
+      required: true
+
+  - type: textarea
+    id: alternatives
+    attributes:
+      label: Alternatives considered
+      description: Workarounds, prior art, or other approaches you considered.
+      placeholder: I currently work around this by manually restarting the provider session, but that loses in-flight context.
+
+  - type: textarea
+    id: tradeoffs
+    attributes:
+      label: Risks or tradeoffs
+      description: What costs, complexity, or edge cases should be considered?
+      placeholder: This may require careful handling when the underlying provider session has already exited.
+
+  - type: textarea
+    id: references
+    attributes:
+      label: Examples or references
+      description: Links, screenshots, mockups, or comparable tools.
+
+  - type: checkboxes
+    id: contribution
+    attributes:
+      label: Contribution
+      options:
+        - label: I would be open to helping implement this.

.github/workflows/issue-labels.yml

@@ -0,0 +1,75 @@
+name: Issue Labels
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - .github/ISSUE_TEMPLATE/**
+      - .github/workflows/issue-labels.yml
+  workflow_dispatch:
+
+permissions:
+  issues: write
+
+jobs:
+  sync:
+    name: Sync issue labels
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Ensure managed issue labels exist
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const managedLabels = [
+              {
+                name: "bug",
+                color: "d73a4a",
+                description: "Something is broken or behaving incorrectly.",
+              },
+              {
+                name: "enhancement",
+                color: "a2eeef",
+                description: "Requested improvement or new capability.",
+              },
+              {
+                name: "needs-triage",
+                color: "fbca04",
+                description: "Issue needs maintainer review and initial categorization.",
+              },
+            ];
+
+            for (const label of managedLabels) {
+              try {
+                const { data: existing } = await github.rest.issues.getLabel({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  name: label.name,
+                });
+
+                if (
+                  existing.color !== label.color ||
+                  (existing.description ?? "") !== label.description
+                ) {
+                  await github.rest.issues.updateLabel({
+                    owner: context.repo.owner,
+                    repo: context.repo.repo,
+                    name: label.name,
+                    color: label.color,
+                    description: label.description,
+                  });
+                }
+              } catch (error) {
+                if (error.status !== 404) {
+                  throw error;
+                }
+
+                await github.rest.issues.createLabel({
+                  owner: context.repo.owner,
+                  repo: context.repo.repo,
+                  name: label.name,
+                  color: label.color,
+                  description: label.description,
+                });
+              }
+            }