fix: Include RetryAfter in AbuseRateLimitError.Error output

[danmoseley]

When AbuseRateLimitError includes a populated RetryAfter field (from the Retry-After HTTP header), the .Error() string silently drops it:

GET https://api.github.com/search/code: 403 You have exceeded a secondary rate limit. Please retry your request again later.

This is inconsistent with RateLimitError.Error(), which already includes timing via formatRateReset:

GET https://api.github.com/search/code: 403 API rate limit exceeded. [rate reset in 2m59s]

Why this matters now

This gap has existed since 2016, but was minor because Go callers can use errors.As to inspect .RetryAfter directly.

It has become more impactful with AI coding agents. The github-mcp-server exposes GitHub API operations as MCP tools, and AI agents receive tool results as plain strings — there is no way for the model to inspect the underlying Go struct. When an agent hits a secondary rate limit, it sees no retry duration and must choose between retrying immediately (worsening the situation) or waiting an arbitrary amount of time.

Including the duration in .Error() gives agent frameworks and models the information they need to back off correctly.

After this change

GET https://api.github.com/search/code: 403 You have exceeded a secondary rate limit. Please retry your request again later. [retry after 1m0s]

When RetryAfter is nil or zero, the output is unchanged.

Fixes #4180.

Note

This PR was authored with the assistance of GitHub Copilot.

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[danmoseley]

@gmlewis

[danmoseley]

hmm, trying to figure out CLA magic

[danmoseley]

OK CLA is good. Ready for review.

[danmoseley]

BTW @gmlewis I wonder whther it might be good in the repo settings to enable copilot code review agent on PR's ? I've gotten used to it. Thats if you didn't try it already and found insufficient value.

[gmlewis]

BTW @gmlewis I wonder whther it might be good in the repo settings to enable copilot code review agent on PR's ? I've gotten used to it. Thats if you didn't try it already and found insufficient value.

I no longer have authorization to modify permissions settings of this repo, @danmoseley.
If the Google Open Source team decides to turn it on, then we will have it available.
Otherwise, people have been using their own access to agents to work on PRs.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.69%. Comparing base (f019b7f) to head (ca9426b).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #4181   +/-   ##
=======================================
  Coverage   93.69%   93.69%           
=======================================
  Files         210      210           
  Lines       19760    19763    +3     
=======================================
+ Hits        18514    18517    +3     
  Misses       1049     1049           
  Partials      197      197           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[gmlewis]

Thank you, @danmoseley!
LGTM.
Awaiting second LGTM+Approval from any other contributor to this repo before merging.

cc: @stevehipwell - @alexandear - @zyfy29 - @Not-Dhananjay-Mishra - @munlicode

[Not-Dhananjay-Mishra]

LGTM

[gmlewis]

Thank you, @Not-Dhananjay-Mishra!
Merging.