fix: stop posting provider errors as failure comments - #82
Merged
Merged
Conversation
A revoked model credential made every dispatch fail at the first model call, and eve's built-in GitHub handlers echoed the provider's own words back into the originating thread: "Model provider API error: Authentication Fails, Your api key: ****53a6...". The failure is an operator problem, not a reply, so it repeated in full on every mention in every conversation. `lib/failure-policy.ts` now decides what a channel may say when a turn or session fails. A deployment fault — an unusable credential, an unpaid gateway account, or a model the account cannot reach — posts nothing and leaves a redacted one-line summary in the runtime log. Every other failure posts one generic sentence plus eve's correlation id, never the upstream provider's text. Both chat channels route their `turn.failed` and `session.failed` events through it. The GitHub channel is where the leak showed, because it had no event overrides at all; `turn.started` stays eve's, since it owns the repository checkout. The Discord mentions channel already posted fixed prose and now suppresses deployment faults too, so a dead credential stops repeating itself there as well.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
This pull request addresses the issue of sensitive provider errors being exposed in GitHub comments. It introduces a new
Reviewers should begin by examining |
1 similar comment
|
This pull request addresses the issue of sensitive provider errors being exposed in GitHub comments. It introduces a new
Reviewers should begin by examining |
This branch was successfully deployed
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.
Closes the noise half of #77.
A revoked model credential fails every dispatch at the first model call, and eve's built-in GitHub failure handlers echoed the provider's own words into the originating thread —
Model provider API error: Authentication Fails, Your api key: ****53a6 is .... The failure is an operator problem, not a reply, and it repeated in full on every mention in every conversation.lib/failure-policy.tsnow decides what a channel may say when a turn or session fails:code,details.name,details.statusCode,details.upstreamType, and the correlation id — never the raw message.Both chat channels route
turn.failedandsession.failedthrough it. The GitHub channel is where the leak showed, because it had no event overrides at all;turn.startedstays eve's own handler, since it owns the repository checkout. The Discord mentions channel already posted fixed prose and now suppresses deployment faults too.node --test lib/*.test.ts(176 tests),tsc --noEmit, andpnpm build:eveall pass.