feat(autofix): surface the running model in every autofix report - #7226
Merged
Conversation
Every visible autofix comment now carries a footer naming Qwen Code and the model it ran — for diagnosis (which model produced this) and as a small attribution for our own model. Four surfaces: the review-address fixed report, the no-action report, the handoff report, and the issue-phase PR's E2E comment. The model comes from the QWEN_PR_REVIEW_MODEL repo variable — already the agent's OPENAI_MODEL, a variable not a secret, so it is safe to echo into a public comment. Each reporting step plumbs it in and computes MODEL_DISPLAY with a 'default' fallback so an unset variable never renders a bare backtick pair. The footer sits with the report body (before the eval marker), and on the E2E path it is appended after the model's file, never injected mid-generation. Contract test pins the env plumbing and the footer on all four surfaces (twice in push-and-report, which carries both bodies), plus the append-after ordering on the E2E path. 61/61 + 12/12.
Collaborator
|
Thanks for the PR! Template looks good ✓ — the "Why it's needed" content is folded into "What this PR does" and the evidence section is adapted for a non-UI change, both fine for a CI workflow PR. Problem: this is a feature addition, not a bug fix — no reproduction needed. The motivation is clear: when an autofix run behaves oddly, there's currently no way to tell which model produced the result. The footer also adds a small attribution for Qwen Code. Both reasons are stated as maintainer-requested. Direction: aligned. This is internal CI tooling that improves observability of the autofix pipeline. No auth, sandbox, model selection, telemetry, release, or public contract concerns. The model name comes from Size: not applicable — no core paths touched ( Approach: the scope feels right. Four reporting surfaces each get the same one-line footer via a shared Moving on to code review. 🔍 中文说明感谢贡献！ 模板完整 ✓ —— "Why it's needed" 内容合并在 "What this PR does" 中，证据部分针对非 UI 变更做了调整，对 CI workflow PR 来说完全合理。 问题：这是功能新增，不是 bug 修复——无需复现。动机清晰：autofix 运行异常时，目前无法判断结果是哪个模型产出的。footer 同时为 Qwen Code 做了一点署名。两个目的均标注为维护者要求。 方向：对齐。这是内部 CI 工具的可观测性改进。不涉及 auth、sandbox、模型选择、遥测、发布或公开契约。模型名取自 规模：不适用——未触及核心路径（仅 方案：范围合理。四个报告位置通过共享的 进入代码审查 🔍 — Qwen Code · qwen3.7-max Reviewed at |
Collaborator
Code ReviewIndependent proposal: given the goal "name the model in every autofix comment", I'd read Comparison with the diff: the PR does exactly this, and does it cleanly.
No critical blockers. No AGENTS.md violations. No security concerns. TestingThis is a CI workflow change — the modified code runs inside GitHub Actions, not in the CLI. There is no user-facing TUI behavior to drive in tmux; the footer only appears when an autofix run completes on GitHub. The appropriate verification is unit tests + static checks: 61/61 autofix tests pass (including the new model-attribution test), 12/12 fleet-shepherd tests pass (untouched), YAML parses, and all modified 中文说明代码审查独立方案： 给定"在每条 autofix 评论中标注模型"的目标，我会将 与 diff 对比： PR 完全按此实现，且很干净。
无关键阻断。无 AGENTS.md 违规。无安全顾虑。 测试这是 CI workflow 变更——修改的代码在 GitHub Actions 中运行，不在 CLI 中。没有用户可见的 TUI 行为可在 tmux 中驱动；footer 只有在 GitHub 上 autofix 运行完成时才出现。合适的验证是单元测试 + 静态检查： 61/61 autofix 测试通过（含新的模型标注测试），12/12 fleet-shepherd 测试通过（未改动），YAML 可解析，所有修改的 — Qwen Code · qwen3.7-max Reviewed at |
Collaborator
|
Confidence: 5/5 — clean across every stage; would merge without hesitation. This is exactly the kind of PR I like to see: small, focused, and solving a real operational need. The diff does one thing — names the model in autofix comments — and does it in the most straightforward way possible. My independent proposal and the actual implementation are essentially identical, which means there wasn't a simpler path the author missed. The If I had to maintain this in six months, I'd thank the author — the footer is self-documenting, the env var comment explains why it's safe to echo, and the test would catch any surface that loses the plumbing. No reservations. Shipping it. ✅ 中文说明置信度：5/5 —— 每个阶段都干净；毫不犹豫地合入。 这正是我喜欢看到的 PR：小、聚焦、解决真实的运维需求。diff 只做一件事——在 autofix 评论中标注模型——而且用最直接的方式完成。我的独立方案和实际实现基本一致，说明作者没有遗漏更简路径。
如果六个月后我来维护，我会感谢作者——footer 自解释，env 变量注释说明了为何可以安全回显，测试会捕获任何丢失注入的位置。 无保留意见。合入。 ✅ — Qwen Code · qwen3.7-max Reviewed at |
qwen-code-ci-bot
previously approved these changes
Jul 19, 2026
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
LGTM, looks ready to ship. ✅
Contributor
Code Coverage Summary
CLI Package - Full Text ReportCore Package - Full Text ReportFor detailed HTML reports, please see the 'coverage-reports-22.x-ubuntu-latest' artifact from the main CI run. |
wenshao
commented
Jul 19, 2026
wenshao
left a comment
wenshao
left a comment
Collaborator
Author
There was a problem hiding this comment.
— qwen3.8-max-preview via Qwen Code /review
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
Reviewed — no blockers. Suggestions are inline.
— qwen3.8-max-preview via Qwen Code /review
Co-authored-by: qwen-code-ci-bot <qwen-code-ci@service.alibaba.com>
qwen-code-ci-bot
approved these changes
Jul 19, 2026
qwen-code-ci-bot
left a comment
qwen-code-ci-bot
left a comment
Collaborator
There was a problem hiding this comment.
No issues found. LGTM! ✅
— qwen3.8-max-preview via Qwen Code /review
Collaborator
|
LGTM — additive display text, no behavior change. Confirmed the parts worth checking: |
Collaborator
|
Released in v0.20.1. |
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
3 participants
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 this PR does
Adds a one-line footer to every visible autofix comment, naming Qwen Code and the model it ran:
Four surfaces: the review-address fixed report, the no-action report, the handoff report, and the issue-phase PR's E2E comment.
Two reasons, both requested by the maintainer: diagnosis (which model produced a given result — invaluable when a run behaves oddly) and a small attribution for our own model.
How it works
The model comes from the
QWEN_PR_REVIEW_MODELrepo variable — already the agent'sOPENAI_MODEL, a variable (not a secret), so it is safe to echo into a public comment. Each reporting step plumbs it into its env and computesMODEL_DISPLAY="${MODEL:-default}", so an unset variable rendersdefaultrather than a bare backtick pair. The footer sits with the workflow-authored report body (immediately before the invisible<!-- autofix-eval -->marker); on the E2E path it is appended to the model-authored file after generation, never injected mid-stream.Reviewer Test Plan
How to verify
npx vitest run scripts/tests/qwen-autofix-workflow.test.js— 61/61. The new test pins theMODELenv plumbing and the footer on all four surfaces (it appears twice in push-and-report, which carries both the fixed and no-action bodies), thedefaultfallback, and the append-after ordering on the E2E path.npx vitest run scripts/tests/qwen-fleet-shepherd-workflow.test.js— 12/12 (untouched).run:block passesbash -n; the footer's escaped backticks render as inline code (verified by executing the body).Evidence (rendered footer)
Tested on
Risk & Scope
QWEN_PR_REVIEW_MODELis a non-secret repo variable already exposed asOPENAI_MODELand in logs; echoing it into a comment leaks nothing new. Purely additive display text.Linked Issues
Independent of, but best seen after, #7225 (P0 review-address fix).
中文说明
本 PR 做了什么
给每一条可见的 autofix 评论加一行 footer,标注 Qwen Code 与本次运行的模型:
四个出现位置:review-address 的已修复报告、无需改动报告、**交接(handoff)**报告,以及 issue 阶段新建 PR 的 E2E 评论。
两个目的(均为维护者要求):诊断(某次结果是哪个模型产出的 —— 运行异常时极有价值)与给自家模型做一点署名/广告。
实现方式
模型取自
QWEN_PR_REVIEW_MODEL仓库变量 —— 本就是 agent 的OPENAI_MODEL,是变量而非 secret,写入公开评论无泄露风险。每个报告步骤把它注入 env 并计算MODEL_DISPLAY="${MODEL:-default}",变量为空时显示default而非空的反引号对。footer 位于 workflow 生成的报告正文中(紧邻不可见的<!-- autofix-eval -->marker 之前);E2E 路径上则在模型生成文件之后追加,绝不中途注入。评审验证
npx vitest run scripts/tests/qwen-autofix-workflow.test.js—— 61/61。新测试钉住四处的MODELenv 注入与 footer(在同时承载「已修复 + 无需改动」两个正文的 push-and-report 中出现两次)、default兜底,以及 E2E 路径的「先生成后追加」顺序。npx vitest run scripts/tests/qwen-fleet-shepherd-workflow.test.js—— 12/12(未改动)。run:块通过bash -n;footer 的转义反引号渲染为行内代码(已实跑正文验证)。风险与范围
QWEN_PR_REVIEW_MODEL是非 secret 仓库变量,本就作为OPENAI_MODEL暴露、也出现在日志中,写入评论不新增泄露。纯展示文本、只增不改。关联 Issue
与 #7225(P0 review-address 修复)相互独立,但建议在其合入后查看效果。