The host-side prep (clone + git diff in the sandbox) was producing empty
diffs for some PRs. The agent then silently published a "no issues
found" review (e.g. langchainplus#24069, #24458). Fix is to rip out the
prep entirely and instruct the agent to fetch the diff itself with
\`gh pr diff <num> --repo <owner>/<repo>\` (or the compare API on
re-review). \`add_finding\` line-range validation is skipped when no
\`diff_line_set\` is provided in config — we trust the agent's anchors.
We can revisit a host-side diff path once we understand why the prep
was failing; this hotfix unblocks reviews on private PRs in the
meantime.
When `_ensure_repo_checked_out` or `compute_diff_in_sandbox` fail, the
reviewer used to log+continue and hand the agent an empty diff. The
agent would then emit a misleading "no issues found" review and the
underlying error stayed buried in server logs.
Now the helpers raise on non-zero exit codes (with the sandbox output)
and the prep block re-raises, so the failure shows up in the LangSmith
run trace and we can debug the real issue.
The reviewer agent was writing a 1–2 sentence "top-level take" as the
review body, which produced noisy paragraph-style summaries on PRs
("Reviewed the PR. The new ALLOWED_GITHUB_REPOS allowlist…"). Devin's
review comment is just a one-liner (`✅ No Issues Found` or
`**Devin Review** found N potential issue.`), and was preferred in the
internal A/B vs Graphite.
Drop the `summary` parameter from `publish_review`; render a fixed,
host-formatted body in `render_review_body` instead. Update the reviewer
prompt to forbid prose summaries.
* fix(reviewer): log every push/close early-return so 'silent ignore' is debuggable
Pushes to PRs that haven't had a first review fall through the watch
handler because the reviewer thread doesn't have kind=reviewer set.
Without log lines on the early-return paths, this scenario was
indistinguishable from 'webhook reached the handler at all' in the
hosted log stream.
Now every early-return logs at info or debug:
- info when a real PR exists but the reviewer thread isn't set up
(with a hint pointing at the trigger paths the user can use)
- info when the repo isn't in the reviewer allowlist
- debug for benign skips (non-branch refs, branch deletions,
already-reviewed head_sha)
* fix(reviewer): always post a summary review, even with no findings
The publish_review tool gated POSTing on `inline_comments or summary`,
so when the agent called publish_review() with no args on a clean PR
the result returned `success: true` but no GitHub review was posted —
the user got silence instead of a "no issues found" comment.
- Drop the gate so publish_review always POSTs.
- Friendlier no-findings render: `**No issues found.**` when the
findings list is empty, vs. `**No issues at or above \`<sev>\`
severity.**` with hidden count when only sub-threshold findings
exist. Agent summary renders below.
- Prompt now requires the agent to always pass a `summary` so the
body is meaningful; calls out specifically not to skip on a clean PR.
* feat: implement reviewer findings, publish_review, and watch mode
Build out the reviewer agent end-to-end against the design in
REVIEWER_DESIGN.md:
- Findings as first-class state on the reviewer thread metadata
(`agent/reviewer_findings.py`): Finding TypedDict with start_line/end_line
ranges, suggestion text for ```suggestion blocks, github_review_comment_id
for cross-run reconciliation, diff_hunk for UI rendering. Thread-level
metadata gets `kind=reviewer`, `pr`, `last_reviewed_sha`, `watch` so a
future frontend can list reviewer threads via the langgraph SDK.
- Diff utilities (`agent/reviewer_diff.py`): parse_unified_diff,
compute_diff_line_set for in-diff validation, extract_diff_hunk for
caching the hunk on a Finding, compute_diff_in_sandbox for SHA-to-SHA
diffs against the prepped repo.
- Tools: `add_finding` (validates against the diff line set so out-of-diff
ranges fail at creation, not at GitHub-publish), `update_finding`,
`list_findings`, `publish_review`. The reviewer agent's tool list is
swapped from `[]` (direct shell `gh api` calls) to these four.
- Publish path (`agent/reviewer_publish.py` + `agent/tools/publish_review.py`):
one POST /reviews call with body + inline comments + ```suggestion blocks,
per-comment IDs stored back on findings, GraphQL `resolveReviewThread`
fired for findings transitioning open->resolved on a re-review.
- Reviewer graph: deterministic clone-or-fetch + checkout in the factory
before the agent's first model call (warm- and cold-path symmetric);
computed diff and in-diff line set passed via runnable config; system
prompt rewritten for the single-evolving-findings model, severity ladder,
in-diff-only discipline, and watch-mode reconciliation flow.
- Watch mode in webapp.py: `push` event + `pull_request` closed/reopened
added to supported events. New `process_github_push_event` resolves the
open PR for the pushed branch, gates on the reviewer thread's `watch`
flag, builds a re-review configurable, and triggers a run on the same
canonical thread. `process_github_pr_close` toggles watch on
closed/reopened. `set_reviewer_thread_metadata` is called on first
review to install `kind=reviewer` + PR identity + watch=True.
- Eval harness: target.py now extracts `add_finding` calls (mapped to the
legacy {file, line, body, severity} shape the judge expects) and passes
the right configurable so the prep step has base/head SHAs.
- Tests: new unit suites for findings helpers, diff parsing, finding tools,
publish rendering + GraphQL resolve, and watch-mode webhook handlers
(push triggers re-review only when watching, idempotent on unchanged
head SHA, PR close disables watch). Updated existing reviewer-webhook
tests to mock `set_reviewer_thread_metadata`.
- REVIEWER_EVAL_PLAN.md removed per user request; folded relevant context
into REVIEWER_DESIGN.md.
* fix(reviewer): correct git diff flags, scope, dedup, and review-comments URL
Address PR #1253 review findings:
- compute_diff_in_sandbox dropped the invalid `--no-prefix=false` flag
(`option no-prefix takes no value` — every prep run was failing
silently and the agent saw an empty diff).
- compute_diff_in_sandbox grew a `merge_base` flag. First-review path
now uses three-dot `base...head` (the merge-base diff GitHub renders
on Files-changed) so we don't pick up changes that landed on the base
branch after the PR diverged. Re-review delta keeps two-dot
`last_reviewed_sha..head` since that's exactly the new commits.
- publish_review skips findings that already carry
`github_review_comment_id`. Without this, watched re-reviews
re-posted every previously surfaced finding, and only the most-recent
duplicate's id would later resolve when the issue got addressed.
- fetch_review_comments URL now includes `{pull_number}` —
`/repos/{owner}/{repo}/pulls/{pr_number}/reviews/{review_id}/comments`
is the canonical endpoint; the old form 404s, so comment ids were
never stored and watch-mode resolution couldn't run.
Three new tests cover: three-dot vs two-dot wiring, no `--no-prefix`
flag in the executed command, and that publish_review does not re-post
findings whose `github_review_comment_id` is set.
* fix(reviewer): default publish cap from 15 to 4
A clean PR with one critical issue padded out by three lower-severity
findings is fine; fifteen is review spam. The agent can override per
call when a PR genuinely warrants more.
* Use gh for reviewer inline comments
* Trigger reviewer on PR review requests
* Format reviewer webhook test
* Add GitHub repo allowlist
* Separate reviewer repo allowlist
* only implement allowlist for reviewer
* feat: add reviewer graph + eval target wiring
- New `reviewer` graph (`agent/reviewer.py`) registered in langgraph.json
alongside the main `agent` graph. Reuses the same sandbox lifecycle,
GH proxy auth, and middleware primitives from `agent.server`, but with
a narrower tool set, a reviewer-specific system prompt, no
commit/push, and the `task` (subagent) tool stripped via
`_ToolExclusionMiddleware` so review stays in one context.
- New `github_comment` tool: agents call it once per issue with
`(file, line, body, severity)` and the eval scores those calls
against golden comments.
- `ensure_no_empty_msg` middleware (the no_op nudge) is intentionally
*not* on the reviewer's stack — that middleware exists to enforce the
main agent's "always finalize via Slack/Linear/PR" contract, which
the reviewer doesn't have. The main agent's behavior is unchanged.
- `evals/reviewer/target.py`: send PR info as a user message, extract
every `github_comment` tool call (multiple expected per review) into
the run output.
- `evals/reviewer/judge.py`: per-example evaluator now returns a list
of metrics under `{"results": [...]}` so LangSmith averages each
numeric key (f1/precision/recall/tp/fp/fn) across the experiment in
the UI. Dropped the broken `aggregate_pr` summary evaluator that
reached for an attribute that doesn't exist on `RunTree`.
- `evals/reviewer/run_eval.py`: `--limit` now slices the dataset via
`client.list_examples(limit=N)` since `aevaluate` doesn't accept
`max_examples`.
- Makefile: `dev` and `run` targets now use `uv run` so they work
without an activated venv.
* resolve comments
---------
Co-authored-by: open-swe[bot] <open-swe@users.noreply.github.com>
