Agent-loop e2e suite for style_update, a11y_fix, error_fix (ANN-6)#37
Open
kurtstohrer wants to merge 1 commit into
Open
Agent-loop e2e suite for style_update, a11y_fix, error_fix (ANN-6)#37kurtstohrer wants to merge 1 commit into
kurtstohrer wants to merge 1 commit into
Conversation
Stands up the agent round-trip plumbing as a measurable, CI-blocking artifact for the public demo and design-partner pitch. Each task type drives a full lifecycle on react-workflows (React+Vite) and vue-data-lab (Vue+Vite): - Capture the AgentLoopTarget component + tracer stylesheet - Seed a deterministic task shape via the per-MFE API - Open the host shell, exercise the task panel / a11y scan / error monitor as appropriate - Run a rule-based simulator that follows the same MCP-CLI sequence (annotask task / update-task) a real coding agent would - Verify the iframe DOM, axe rescan, or console stream reflects the fix after Vite HMR - Restore the captured files and emit a per-run JSON metric The simulator is intentionally rule-based for v1 — see docs/agent-loop-evals.md for the schema, the caveats around what is and isn't measured today, and where the LLM apply step plugs in for v2. Per-test metrics land under playgrounds/stress-test/e2e/annotask/reports/agent-loop/ and are uploaded as a CI artifact. A focused playwright config (agent-loop.config.ts) only spins up host + the two target MFEs so the new CI job stays under the broader stress-cluster cost. The existing pnpm test:e2e:stress:annotask script still picks the tests up via its directory filter. Co-Authored-By: Paperclip <noreply@paperclip.ing>
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.
Summary
style_update,a11y_fix,error_fix— run on React+Vite (react-workflows) and Vue+Vite (vue-data-lab) MFEs in the stress lab.AgentLoopTargetcomponent, seeds a deterministic task, drives the shell, runs a rule-based simulator that follows the same MCP-CLI sequence a real coding agent would, verifies the iframe DOM / axe rescan / console stream after Vite HMR, then restores the captured files.playgrounds/stress-test/e2e/annotask/reports/agent-loop/. Schema documented indocs/agent-loop-evals.md; CI uploads the directory as an artifact.v1 caveat
The simulator's apply step is intentionally rule-based, not LLM-driven. The harness measures plumbing reliability (does the task land, do the MCP tools work, does HMR pick the fix up, do metrics persist). LLM apply quality is the follow-up ticket — see the doc for where the LLM plugs into the same harness.
Test plan
agent-loopjob runspnpm test:e2e:stress:annotask:agent-loopagainst the focused playwright config and uploads the metrics artifactpnpm test:e2e:stress:annotaskstill picks the new specs up via itsannotask/directory filterpnpm build && pnpm test:e2e:stress:annotask:agent-looplocally with the stress host + react-workflows + vue-data-lab dev servers reachabledocs/agent-loop-evals.mdmatchesCloses ANN-6.