Release: merge main into releases/v1#636
Conversation
The result comment links to app.meticulous.ai, which coding agents reading the comment via the GitHub API can't access directly. Append an HTML-comment block (invisible in the rendered GitHub UI, but present in the raw body) pointing agents to the CLI/MCP server and the meticulous-review skill instead. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
## Summary
- The Meticulous result comment on a PR links to `app.meticulous.ai`,
which coding agents reading the comment via `gh api`/REST/GraphQL can't
open directly (no browser session).
- Appends a hidden HTML-comment block to the result comment (invisible
in the rendered GitHub UI, but present in the raw comment body) that
tells an agent how to fetch the same results via the Meticulous CLI or
MCP server, and points it at the `meticulous-review` skill for how to
approach the review.
- Only added to the three `testRunFinished` branches that actually
include a result URL.
## Test plan
- [x] `tsc --noEmit` passes
- [x] `eslint` passes
- [x] CLI command names/flags cross-checked against
`~/meticulous/public_packages/cli/src/commands/agent/*` (e.g.
`--replayDiffId`, not `--replayDiff`)
- [ ] Manually verify the rendered PR comment on a real test PR still
looks correct and the hidden block is retrievable via `gh api
repos/{owner}/{repo}/issues/comments/{id}`
🤖 Generated with [Claude Code](https://claude.com/claude-code)
<!-- CURSOR_SUMMARY -->
---
> [!NOTE]
> **Low Risk**
> Comment-body-only change in CI reporting; no auth, data, or test
execution logic affected.
>
> **Overview**
> **Meticulous GitHub result comments** now include a hidden
HTML-comment block (visible in raw API bodies, not in rendered PR UI) so
coding agents can reach the same test-run data without opening
`app.meticulous.ai`.
>
> `ResultsReporter` gains `buildAgentAccessNote(testRunId)`, which
documents CLI/MCP setup, the `meticulous-review` skill, and commands to
list diffs and fetch images/DOM/timeline for a run. That note is
appended only on the three `testRunFinished` status comments that
include a results URL (zero diffs, first-run/no baseline compare, and
visual diffs). Comments without a result link (e.g. replay errors with
no snapshots) are unchanged.
>
> <sup>Reviewed by [Cursor Bugbot](https://cursor.com/bugbot) for commit
1b02172. Bugbot is set up for automated
code reviews on this repo. Configure
[here](https://www.cursor.com/dashboard/bugbot).</sup>
<!-- /CURSOR_SUMMARY -->
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
Test suite: ✅ Meticulous spotted 0 visual differences across 9 screens tested: view results. Expected differences? Click here. Last updated for commit |
|
✅ Meticulous spotted zero visual differences across 9 screens tested: view results. Test suite: test Meticulous with deployment url. Last updated for commit 6e3b427. This comment will update as new commits are pushed. |
|
✅ Meticulous spotted zero visual differences across 9 screens tested: view results. Test suite: test Meticulous with app url. Last updated for commit 6e3b427. This comment will update as new commits are pushed. |
Merging
mainintoreleases/v1ahead of cutting a newv1.x.yrelease.Includes:
🤖 Generated with Claude Code
Note
Low Risk
Comment-body-only change for GitHub status updates; no auth, data, or test execution logic is modified.
Overview
Meticulous GitHub PR status comments now append a machine-readable note (via an HTML comment) that tells coding agents how to fetch test-run diffs through the CLI or MCP, including the concrete
testRunIdfor the run.The note is wired into the main test-run finished outcomes: zero visual differences, visual differences detected, and the “no baseline on base branch” first-run message. Human-facing comment text is unchanged in the rendered GitHub UI; agents reading the raw comment body via the API still see the guidance.
Reviewed by Cursor Bugbot for commit 6e3b427. Bugbot is set up for automated code reviews on this repo. Configure here.