feat(error-tracking): add the error-tracking program on the orchestrator flow - #1185
Draft
ablaszkiewicz wants to merge 5 commits into
Draft
feat(error-tracking): add the error-tracking program on the orchestrator flow#1185ablaszkiewicz wants to merge 5 commits into
ablaszkiewicz wants to merge 5 commits into
Conversation
…tor flow New `wizard error-tracking` command. Modeled on replay-vision: a detect step carries the framework id into orchestrator preflight, the context-mill `error-tracking` flow seeds install/init (integration-v2 step-skills) when the repo has no PostHog, then exception capture, then the source-map subgraph where the platform needs it. The detect step also pre-installs posthog-cli for symbol-upload platforms, which the agent cannot (warlock blocks npm install -g). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY
🧙 Wizard CIRun the Wizard CI and test your changes against wizard-workbench example apps by replying with a GitHub comment using one of the following commands: Test all apps:
Test all apps in a directory:
Test an individual app:
Show more apps
Test against a Context Mill branch:
Add Results will be posted here when complete. |
This was referenced Sep 1, 2026
Binding harness anthropic→pi, like metrics: the binding routes only, and every stage's model and effort are pinned context-mill side in the flow's frontmatter (terra seed, sol tasks, luna report). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY
e2e.json flow definition + profile registration: the driver confirms the intro, answers the flow's api-key ask from SOURCE_MAPS_CLI_KEY (secret rule, vaulted by the ask tool) and declines the test-affordance offer. Two variations: the pinned orchestrator-on-pi binding and the linear fallback. Unlocks snapshots mode and the workbench e2e runner for the command. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY
The screen was registered in the action registry (satisfying the exhaustiveness test) but decideE2eAction keeps its own intro switch, and an unlisted screen falls through to wait — the driver sat on the intro until the host died. Adds the missing case. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY
The bin resolves --local-* flags in its yargs middleware into the local-dev singleton that getSkillsBaseUrl() reads. The tui-host bypasses yargs, so its env-backed local flags reached the session but never the singleton — agent prompts always resolved to production, and an unpublished flow died in preflight with "No seed agent prompt". Initialize the singleton from the same env spellings before the store starts. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Problem
Setting up PostHog error tracking today takes multiple commands: the default integration flow, then
wizard upload-source-maps— and the source-maps command aborts on repos without PostHog instead of installing it. Replay-vision already solved this shape (install PostHog as step 0 via the orchestrator seed); error tracking deserves the same one-command experience.Changes
New
wizard error-trackingcommand, modeled on replay-vision:src/lib/programs/error-tracking/): detect step puts the framework id onsession.skillIdfor orchestrator preflight,agentFlow: 'error-tracking'pinned,ciPreRunfor headless runs. No platform allow-list — every detectable framework has anintegration-v2-error-tracking-stepvariant.posthog-clifor symbol-upload platforms (swift, android, react-native, flutter, go, rust) — the agent can't, warlock blocksnpm install -g. Warn-don't-fail, mirroring the source-maps program.error_tracking:read,product_enablement:write), custom intro screen (no single skillId exists for the generic intro), command wrapper +bin.ts, e2e action registry entry.context/agents/error-tracking/): seed checks whether PostHog is integrated and which uploader variant applies, then queues install/init → capture-exceptions → the source-map subgraph (credentials/configure/wire-ci/test-setup) → report. Companion PRs in context-mill and wizard-workbench.wizard upload-source-mapsstays as-is — it remains a valid standalone entry point.Test plan
pnpm build && pnpm test && pnpm lint— 2586/2586 tests pass, including newerror-tracking.test.tsand the updated lockstep guards (switchboard bindings, flags isolation, e2e action registry exhaustiveness).--local-context-millvia the workbench stack); not yet run against a live app.LLM context
Built with Claude Code; flow design mirrors
replay-vision(wizard) anderror-tracking-upload-source-maps(context-mill agents).🤖 Generated with Claude Code
https://claude.ai/code/session_01PVXmLWB31CBujD2FJ2ogvY