fix(evi): have first responder reply directly instead of narrating a draft - #572
Merged
Conversation
|
Contributor
|
The latest updates on your projects. Learn more about Vercel for GitHub.
4 Skipped Deployments
|
Contributor
|
Thank you for following the naming conventions! 🙏 |
commit: |
HugoRCD
approved these changes
Aug 14, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Rewrites the first-responder instruction so the reply reads as a direct comment from Evi to the reporter, not as a meta "draft for the channel".
Why
On issue #571 the posted comment led with "Labels
bugandadapterare applied. I can't post the comment directly on this autonomous turn, so here's the draft reply for the channel:" and a "Draft comment for #571" banner. The cause is in the instruction itself, which said: "Keep the reply to one comment, the draft response Hugo builds on… the channel posts the comment." That phrasing told the model it was producing a draft and narrating the process, so the channel then posted that narration verbatim as the comment.The new text makes it explicit that the single reply the model writes is the comment the channel posts, that it is addressed to the reporter, and that it must never describe drafts, approval, or channel mechanics.
No authorization change
Label tools are already authorized for autonomous turns (
addLabels: autonomousWrite,createLabel: createLabelPolicyinapps/evi/agent/extensions/github.ts); that is whybug+adapterapplied correctly on #571. The first-responder eval still pins the design: the reply is posted by the channel andgithub__addIssueCommentstays out of the autonomous toolset, so no permission or eval change is needed.Checks
pnpm --filter evi run typecheck: exit 0pnpm run lint: 22 tasks successful, exit 0apps/evi(instruction text only), so no changeset is required.