pi-review

Pi extension for code review that runs in the foreground: /review hands a directive to the main agent, which fans out parallel reviewers + a cheap-model gate via the subagent tool. The whole review streams in chat — no silent background work.

Requires the pi-subagents extension (pi install npm:@mariozechner/pi-subagents) — it provides the subagent tool used to fan out.

Pattern ported from the Claude code-review plugin. See reference/ for upstream flow notes and the version roadmap.

Commands

/review [any prompt] [--lite] [--gate-model <id>]
/review-config
/review-agents

/review

CC-aligned: text after /review is user context. The main agent obtains the diff once into .pi/pi-review/change.diff; lean pi-review.* subagents review that file (same orchestration idea as Claude /code-review, with token budgets pinned in the directive).

When called with no arguments, /review targets local git and tells the main agent to:

1. If the working tree is dirty → git diff HEAD into .pi/pi-review/change.diff.
2. If clean → git fetch origin <default-branch>, then git diff origin/<default-branch>...HEAD (avoids a stale local main/master).
3. Also writes changed-files.txt, change-kind.txt (docs|code), and diff-meta.txt (base/head SHAs).
4. Outside a git repo with no PR → notify and exit.

PR review: pass a GitHub PR URL or number. The main agent tries gh pr diff; if the diff is too large (GitHub 20k-line limit) or unavailable, it fetches pull/<n>/head + the PR base ref and runs a three-dot git diff. Oversized PRs no longer fail in the plugin layer.

/review https://github.com/org/repo/pull/17206
/review 17206
/review focus on backup restore — PR 17206

Lite mode: add --lite for a fast single-agent pass — one reviewer covers all dimensions (bugs / security / compliance / code-comments / history), no fan-out, no gate. Lower latency and cost; ideal for a quick gut-check.

/review --lite
/review --lite focus on concurrency and secret handling

Flags

The surface is intentionally minimal. Removed knobs (--threshold / --reviewer / --no-gate / --score-per-issue / --diff) are accepted-but-ignored so old invocations do not break; their capabilities now live in config.json via /review-config.

Flag	Effect
--lite	Single-agent fast mode: one reviewer, no gate.
--gate-model <id>	Override the gate model for this run (otherwise config.gate.model).
--no-spawn	Dry run — print the directive that would be injected, and exit.

Removed flags → config

Old flag	config.json key
--threshold N	gate.threshold
--reviewer id	reviewers.<id>.enabled
--no-gate	gate.enabled
--score-per-issue MODE	gate.scorePerIssue
--diff path	(removed — pass a PR url or run in a dirty repo instead)

Bundled reviewers (v0.5.1 lean agents)

Runtime names are pi-review.<id> (package agents registered via pi.subagents.agents).

ID	Purpose	Default	Tools
claude-md-compliance	Project rules (AGENTS.md / CLAUDE.md / .pi/)	enabled	read, grep, ls
bugbot	Obvious bugs in introduced lines only	enabled	read, grep, bash
history-context	Light git blame / log (≤5 files, log -n 5)	enabled	read, bash
security-review	Security issues introduced by the change	enabled	read, grep, bash
code-comments	Inline comment / TODO guidance in changed files	enabled	read, grep
conventions	De-facto style pass	disabled	read, grep, ls
gate	Dedupe + re-score + verdict	(always)	read
lite-review	Single-agent all-dimensions (--lite)	on demand	read, grep, bash

Pipeline

/review runs in the foreground: the handler builds a directive and injects it hidden into the main agent. The main agent obtains the diff once, then fans out lean reviewers via subagent (exactly one parallel call + exactly one gate):

Step 1  obtain → .pi/pi-review/change.diff + changed-files.txt + change-kind.txt
Step 2  subagent({ tasks: [pi-review.*…], turnBudget≈20, file-only })  — once
Step 3  subagent({ agent: pi-review.gate, model: <gate>:<thinking> }) — once
Step 4  report from file-only outputs (do not re-read the diff)

Permissions: on each /review, CC-aligned allow rules (7× gh + read-only git + Read/Grep) are merged into .pi/projects/<id>/permissions.local.json so headless reviewers are not blocked by permission-modes (no ask UI in children).

Cost: dominated by N × tool turns. Prefer --lite for a cheap pass. Reviewer thinking inherits the parent session; only the gate uses config.gate.thinking.

Configuration

Per-user config lives at:

~/.pi/agent/pi-review.json

Run /review-config to open it in $EDITOR. The file is loaded, merged with the built-in defaults, and validated on every /review call. Unknown reviewer ids are added as new reviewers; known ids are patched in place.

{
  "schemaVersion": 1,
  "gate": {
    // Cheap tier by default (pure de-noise reasoning). "inherit" follows the parent session.
    // Override to a specific id; non-anthropic providers must set this explicitly.
    "model": "anthropic/claude-haiku-4-5",
    "thinking": "low",
    "enabled": true,
    // Issues with confidence < threshold are dropped (code-enforced).
    "threshold": 8,
    // Per-issue scorers: "off" (default) | "blocker-major" | "all"
    "scorePerIssue": "off"
  },
  "concurrency": 4,
  "budgets": {
    "turnBudget": { "maxTurns": 20, "graceTurns": 2 }
  },
  "reviewers": {
    "claude-md-compliance": {
      "model": "anthropic/claude-opus-4-6",
      "thinking": "high",
      "enabled": true
    },
    "bugbot": {
      "model": "inherit",
      "thinking": "medium"
      // tools inherit from inheritance.toolsDefault
    }
  },
  "inheritance": {
    "toolsDefault": ["read", "grep", "find", "ls", "bash"],
    "inheritProjectContext": false,
    "inheritSkills": false
  }
}

Use "model": "inherit" on a reviewer to follow the parent session's model. The gate defaults to a cheap tier; override it persistently via gate.model in config, or per-run with --gate-model <id> (the directive injects it into the gate subagent).

TUI output

The main agent writes the report directly into chat. Shape (illustrative):

## pi-review — uncommitted changes

**Verdict: REQUEST_CHANGES** (1 blockers · 1 major · 0 minor · 0 nit)
Reviewed in 42.1s · 4 reviewers · 1 gate

### claude-md-compliance (anthropic/claude-opus-4-6) — ok · 12.0s
- 1 issue
- [BLOCKER · compliance · conf 9] `src/auth.ts:42` — rule "no raw token logging" violated

### bug-detector (anthropic/claude-sonnet-4-6) — ok · 8.1s
- 1 issue
- [MAJOR · bug · conf 8] `src/auth.ts:42` — race: missing await on fetchUser

### conventions (anthropic/claude-sonnet-4-6) — ok · 6.2s
- 0 issues
- all naming/imports match surrounding files

### history-context (anthropic/claude-haiku-4-5) — ok · 4.8s
- 0 issues
- touched files have stable history, no recent reverts

### gate (anthropic/claude-haiku-4-5) — ok · 5.4s
- verdict: request_changes
- reason: One auth.ts blocker + one race condition in worker.ts; safe to merge after fixes.
- 2 issues after dedupe + threshold

A machine-readable copy is also written via pi.appendEntry("pi-review", { ... }) for future TUI consumers.

Install

pi install npm:@georgedong32/pi-review

Verify:

pi list

Local development

bun install
bun run check     # tsc --noEmit
bun test          # node:test + tsx

Repo structure

index.ts                  Pi extension entry; registers /review + /review-config + /review-agents
src/types.ts              Shared interfaces (Issue, ReviewerSpec, PiReviewConfig, ReviewReport)
src/config.ts             loadConfig / mergeWithDefaults / validateConfig / writeConfig
src/args.ts               buildReviewerArgs / buildGateArgs / applyThinkingSuffix
src/spawn.ts              runSubagent (child_process.spawn + structured-output read)
src/schema.ts             TypeBox schemas for reviewer + gate outputs
src/parallel.ts           mapConcurrent (hard cap 4)
src/review.ts             runReviewers — fan-out + per-reviewer spawn
src/gate.ts               runGate — single spawn with aggregated prompt
src/git-input.ts          resolveDefaultDiff (status / vs default branch)
src/report.ts             buildReport + renderReport
agents/*.md               Bundled reviewer prompts (incl. lite-review.md for --lite)
prompts/gate.md           Gate subagent prompt
tests/*.test.ts           node:test suites

Limitations (v1)

• No retry: a failed reviewer is recorded as ok=false and the rest of the run continues. The gate still runs.
• No worktree isolation: all reviewers share the parent's cwd.
• GitHub: agents use gh/git to obtain PRs; oversized diffs fall back to git. PR comments are not posted.
• No web config UI: only $EDITOR.

License

Apache-2.0