review-pr

name	review-pr
description	Autonomous PR review that posts high-signal, human-voiced comments under your account. Use when reviewing someone else's PR or your own manifest-driven PR, when you want a precision-tuned review you can walk away from, or when the user asks to review a PR, post a PR review, autoreview, loop review, or watch a PR.
argument-hint	[pr-url] [--manifest <path>] [--bundle <urls>] [--loop]
user-invocable	true

High-signal autonomous PR review posted under your account. A review you'd put your name on — precision over coverage.

Inputs. pr-url from the arg or the current branch's upstream PR. --manifest <path> switches the skill into manifest mode: it skips the generic reviewer fleet and independently verifies only the manifest's contract — it does not merely ground the fleet against author intent. Only in that mode, load references/MANIFEST_MODE.md for the verification mechanics; without --manifest, that reference is never read and review runs the generic review-code fleet. The skill does not auto-discover a manifest from any folder convention. --bundle <urls> plus PR-description linked-PR parsing (Depends on #N, Stack:, Co-changes:, GitHub PR URLs) provides cross-PR context for coupled changes. Resolve the PR, current head SHA, our prior GitHub reviews/comments/replies, open review threads, author commits, PR description, and linked-PR context before deciding what to do.

Self-marking. Posting under your own account means a human reviewer uses the same account, so authorship alone can't tell our automated comments from theirs. Stamp every body this skill posts — new finding comments, thread replies, the summary header, the approval body, and manifest-mode PASS/FAIL — with a trailing hidden marker <!-- manifest-dev:review-pr --> — a fixed literal string, byte-identical wherever review-pr runs (don't namespace-rewrite it per distribution, or comments posted by one host stop matching another's); GitHub strips it from the rendered comment but returns it through the API. This marker, not account authorship, is what makes a comment an automated review-pr comment: everywhere this skill says threads/comments we authored or replied to (or our prior reviews/comments) it means the ones carrying this marker, so an unmarked comment on our account reads as human and is left untouched. Manifest mode extends the marker with the criterion id (references/MANIFEST_MODE.md). Judgment-pass findings extend it with a judgment token — <!-- manifest-dev:review-pr judgment --> — so a prior judgment finding is distinguishable from an ordinary review-pr comment in both modes (the once-per-PR gate keys on this token). Comments predating this marker are unmarked and so now read as human — review-pr stops advancing its own pre-marker threads, which self-heals as new marked comments accrue.

One-Shot Pass

Every invocation, including non---loop, performs one complete PR-state advance:

1. Advance our existing threads. For every unresolved thread we authored or replied to, run the per-comment verifier below. Post needed thread replies, resolve terminal threads, and leave genuinely pending threads open.
2. Verify the change. Manifest mode (--manifest): load references/MANIFEST_MODE.md and follow it to verify the manifest contract against the PR head — the generic reviewer fleet is skipped entirely. No-manifest mode: run the generic reviewer fleet over the review range — determine that range from durable GitHub state: if we have a prior review on this PR, use that review's commit/head SHA as the lower bound and review last-reviewed-by-us..current-head; otherwise review the full PR diff. In both modes, the judgment pass (below) runs in parallel with the fleet/contract verification, gated to once per PR. In either mode, if the head is unchanged from our latest review, skip re-verifying the code only after thread advancement has run.
3. Post outcomes. Submit new surviving findings as a single GitHub review with decision comment. Thread replies are posted on their existing threads, not as new review comments. End with the cycle summary below.

The one-shot pass is CI-shaped: it must make useful progress from only GitHub state and the current checkout. Do not rely on session memory such as last-reviewed-sha; derive it from our prior review/comment metadata and the PR history each run.

Per-Comment Verifier

For every unresolved thread we authored or replied to, spawn one verifier subagent. The subagent receives:

• The original finding: anchor, body, review commit/head SHA, and concrete concern.
• Every author reply on this thread since our last message, plus every prior reply we posted on this thread.
• Current code at the anchor, with nearby context, and any commits since the original review that touched the relevant range.
• Full PR history: all comments, review threads, commit messages, current PR description, and our prior review/comment bodies.
• Bundle context for each linked PR (≤5): diff, description, top-level conversation.
• The voice profile below for drafting replies.

The subagent returns exactly one disposition:

• addressed-by-fix — A commit after our review modified the relevant code AND removes the concrete failure mode. Resolve the thread. Straightforward code fixes do not need a reply unless there was active discussion to answer.
• addressed-by-valid-reply — The author replied on this or another thread with an argument a fair-minded human reviewer would concede: correct factual context, deliberate owner trade-off, valid out-of-scope boundary, or code elsewhere already covering the concern. Post a short concession/acknowledgment reply, then resolve.
• false-positive-or-stale — Our comment was wrong, stale, or disproven by current PR history. Post a brief correction that owns the miss, then resolve.
• needs-our-pushback — The author replied or changed code, but the concrete concern remains. Post a short reply on the existing thread and keep it pending.
• still-pending — No relevant new author reply and no relevant code change since our last look. Keep the thread pending without posting.

Push back only on new signal, and stay on the specific point under contention. Do not repost the same argument without a new author reply or code change.

Reviewer fleet. No-manifest mode only — skipped entirely in manifest mode (references/MANIFEST_MODE.md). Each lens is a dimension of the manifest-dev:review-code skill — spawn one general-purpose subagent per dimension and have it activate the review-code skill with that dimension against the review range. Always-on dimensions: change-intent, code-bugs, code-design, code-maintainability, code-simplicity, context-file-adherence. Add when the diff fits: type-safety on typed code; test-quality and code-testability on source; contracts on API surfaces; operational-readiness on CI/infra/env/migrations/workers/queues/secrets; docs and prose-value on prose. For prompts/skills/agents, spawn a general-purpose subagent that activates the manifest-dev-tools:review-prompt skill (plus, if available, the external prompt-engineering-plugin agents prompt-token-efficiency-verifier and prompt-compression-verifier). Forward the manifest to every spawned reviewer when present.

Narrow-lens reviewers. Fleet-dimension agents never receive PR conversation, linked-PR diffs, or linked-PR conversation. That context flows only to the holistic pass and to the judgment pass — narrow lens is what keeps each fleet reviewer precise. The judgment pass is the deliberate exception: it is wide-context by design, because premise-questioning without the stated pain would be arrogant guessing.

Holistic coherence pass. Collect findings from the fleet (drop Low severity) and from the parallel judgment pass (below — a distinct non-severity class, not subject to drop-Low), and spawn one subagent with what remains plus:

• PR history: all comments and threads on the PR (including our own from any prior review pass), the author's recent commit messages on the branch, the PR description.
• Bundle context for each linked PR: diff, description, top-level conversation. No inline review comments from linked PRs.
• The manifest if present.
• The reviewed range for this invocation — the code the fleet's defect findings are bounded to.
• The judgment pass's findings, from its parallel wide-context subagent (below). They are consolidated alongside the fleet's but stay exempt from drop-Low and from reviewed-range bounding — the judgment subagent reads the whole PR head itself, so its findings are already whole-PR.
• Any truncation the caller did, carried forward.

The subagent:

• Prunes any finding already covered on the PR: any prior comment (ours from a previous run, or another reviewer's), any concession or rebuttal on an existing thread, anything contradicted by the manifest, a commit message, or the PR description, or piling on an active thread.
• Dedupes across reviewers: merge near-duplicates into one comment when they raise the same underlying concern.
• Premise-subsumption. When a judgment finding questions the necessity or existence of a surface, subordinate the fleet's findings on that same surface under the premise question — drop the low-value ones, and fold any that survive beneath it ("if you keep X: also a, b") rather than listing them as independent peers. Never hard-delete a severe independent defect: the judgment finding is non-blocking, so a defect that matters even if the surface stays is still surfaced on its own.
• Bounds surfaced findings to issues introduced or exposed by the reviewed range, unless this is the first review and the range is the full PR.
• Anchors each surviving finding to exactly one of inline file:line (default), file-level (whole-file concern), or PR-level (cross-cutting, no specific anchor).
• Rewrites every comment body and any drafted reply in the voice profile below.
• Omits a summary header by default — adds one only when there's a real overall take the per-comment list misses (one short sentence, voice-compliant, no boilerplate).

Returns: comments to post (anchor + voice-compliant body), summary header text if any, truncation notes, and a brief dropped-findings tally with dominant reasons.

Judgment pass (premise check). Runs in both modes; in manifest mode it wires via references/MANIFEST_MODE.md. It is its own subagent, spawned in parallel with the fleet (no-manifest) — not folded into the holistic pass, which would mix generative premise analysis into an editorial consolidation job and put it on the critical path. Unlike the narrow-lens fleet it is wide-context: fed the PR description, conversation, codebase direction, and the full PR-head diff (base..head, the whole PR), so its whole-PR altitude holds even on an incremental/loop pass where the reviewed range is only the latest delta. It runs references/JUDGMENT_PASS.md: the whole-PR question of whether the change earns its keep against the pain it solves, asked only when concrete evidence fires a trigger. Its findings are a distinct non-severity class ({trigger, evidence, author-facing question}). The judgment subagent itself synthesizes one question per root before returning (so this holds in both modes, independent of the holistic pass). In no-manifest mode those findings then flow into the holistic coherence pass, which dedupes/merges them alongside fleet findings but exempts them from the drop-Low step (they surface whenever the evidence-gate fires and the PR does not already cover the point, never bounded to the reviewed range) and applies premise-subsumption (see the holistic pass above). They post as author-answerable questions through the same voice, self-marker, and single batched comment review; the judgment pass never blocks, never submits request_changes, and never auto-approves.

Once per PR. Fresh judgment generation runs only when no prior judgment-marked finding (<!-- manifest-dev:review-pr judgment -->) exists on the PR — read from durable GitHub state, never session memory. So once the pass has posted a question, later passes skip regeneration and never re-nag; the discriminator token is what makes "a prior judgment finding exists" detectable (the base marker alone can't — every review-pr comment carries it). This matches its kind: defect-finding re-runs each round on new commits, but a premise question is answered once by the human, so re-posting it every round would be the nag this pass exists to avoid. Advancing an existing judgment thread (reply/resolve via the per-comment verifier) is not gated by this — only fresh generation is. Two honest bounds: (1) a pass that stays silent posts no judgment marker, so on a PR the pass keeps finding sound it may re-run on later rounds — bounded and harmless, since it stays silent (no nag), and the whole-PR read is the same cost the fleet already pays each round; (2) a premise concern introduced by a late commit after a question was already posted is not re-caught — most premise issues are visible from the initial PR, and re-scanning every round is the noise.

Voice. Each comment is one thought: state the problem, point to evidence inline (file:line, short code excerpt when load-bearing), suggest the fix. Direct, concrete, no softeners.

Never in a posted body, header, or thread reply: severity labels ([High], ⚠️, Critical:); emoji of any kind; em-dash rhetorical flourishes ("It's not just X — it's Y" / "not just A, but B"); softeners ("I think", "I recommend", "It seems", "Perhaps consider"); opener boilerplate ("Great PR!", "Nice change, but..."); "at the location above" / "as mentioned" (always name file:line inline); AI disclosure footer.

Structural defaults: prose, not bullets, for a single suggestion; no markdown headers or bold-the-takeaway when the comment is one thought. Headers and bullets are fine when a comment genuinely covers multiple distinct thoughts or parallel items.

Target voice: "Empty input skips the null check — if (input?.value) at parser.ts:42 short-circuits before the parse at parser.ts:47, so {} reaches parse() without the guard. Tighten to if (input?.value != null), or move the parse() call inside the existing branch."

Posting. When the holistic pass returns comments to post, submit a single GitHub PR review with decision comment — all comments batched atomically. In CI or non-interactive contexts, do not wait for approval.

Zero comments to post. When the holistic pass returns nothing and all our existing threads reached terminal disposition, report clean. In interactive sessions only, ask: "Looks good to me. Post as approval on the PR?". Approve → submit decision approve with body Looks good to me.; decline → take no PR action. In CI or non-interactive contexts, take no approval action.

Cycle summary. Every one-shot pass ends with an operator-facing summary, whether it posted comments, resolved threads, asked for approval, or found nothing to do. Keep it compact but complete:

• Reviewed range/head and concrete PR actions taken: new review comment count and anchors, thread replies, resolved threads, approval prompt/action, or no PR action.
• Per-comment verifier subagents: one line per thread naming the anchor, disposition, and the verifier's short reason.
• Reviewer fleet subagents (no-manifest mode): one line per spawned reviewer naming its actionable findings count and the substance of what it found, or none. In manifest mode, report per-criterion verifier results instead, per references/MANIFEST_MODE.md.
• Holistic coherence pass (no-manifest mode): surviving comments, dedupes/merges, pruned findings with dominant reasons, range-bounding decisions, summary header if any, and truncation notes.
• Judgment pass (no-manifest mode): premise questions posted — count and their roots — or none. In manifest mode, report it per references/MANIFEST_MODE.md.

The cycle summary is for the operator transcript or run log only. Do not paste it into PR review bodies, thread replies, or approval text; the only posted summary-like text is the voice-compliant summary header returned by the holistic pass.

Gotchas.

• The only path to decision approve is the user-confirmed lgtm prompt above. Never submit approve automatically anywhere else.
• Never submit decision request_changes — this skill does not algorithmically block merges.
• Never add an AI disclosure footer to a comment, summary, or reply. There is no flag for one.
• Never forward PR conversation or bundle context to a reviewer agent. Only the holistic pass may see that context.
• Never re-raise a finding the holistic pass pruned in this run.
• Never skip thread advancement because the code review range is empty; thread state can change without a new commit.
• In manifest mode (--manifest), follow references/MANIFEST_MODE.md for the verification, posting, and fingerprinting rules and its mode-specific gotchas.

--loop. Load references/LOOP.md.