overnight-review-panel-blocked-reviewer-reads-as-cleanlisted

Overnight specialization of `code-reviewer-subagent-no-bash-blocked-on-pr-diff` (the general tool-gap mechanism). In an UNATTENDED overnight review panel, a reviewer that couldn't see the code reads as a CLEAN one — so a real bug ships by morning. The usual cause: code-review subagents (feature-dev:code-reviewer, voltagent-*, Explore) are frequently provisioned WITHOUT a Bash tool, so when prompted to "review PR #N, fetch the diff with gh pr diff" they return a BLOCKED report (no review performed), or silently review the current checkout (often `main`, which predates the PR). Use when: (1) an overnight review panel (e.g. `overnight-multi-issue-implementation` Phase C, or any `agent-review-panel` run) dispatches reviewers against GitHub PRs or branches not checked out in the working tree; (2) a reviewer returns "I have no shell/gh/git tool" or "the PR sources are not in the working tree"; (3) one reviewer in a parallel panel comes back BLOCKED while siblings succeeded. Fix: pre-generate per-base diffs to files
wan-huiyan/overnight-workflows · ★ 1 · Code & Development · score 71

Install: claude install-skill wan-huiyan/overnight-workflows

# Overnight review panel: a BLOCKED reviewer reads as CLEAN by morning > **Overnight specialization.** The bare mechanism — a code-review subagent with > no `Bash` tool can't `gh pr diff`/checkout a PR, so it returns BLOCKED — is the > general skill `code-reviewer-subagent-no-bash-blocked-on-pr-diff`, applicable to > any PR review. **This** skill is the overnight-specific consequence: in an > unattended run there's no human to notice the BLOCKED report, so it silently > counts as a passing reviewer and the bug ships. ## Problem Every workflow in this repo leans on a **multi-agent review panel** to catch errors before they reach the client (core pattern #1). The panel silently assumes each reviewer can *see the code under review*. It can't always. You dispatch a review subagent (e.g. `feature-dev:code-reviewer`, a `voltagent-*` reviewer, or `Explore`) with a prompt like "use `gh pr diff <N>` to get each PR's diff, then review." The agent returns a **BLOCKED report** — no findings — explaining it has no shell/`gh`/`git` tool and the PR source isn't in the working tree. Meanwhile a sibling reviewer in the same panel may succeed (if it happened to find a materialized worktree on disk). In an interactive session you'd notice. **In an unattended overnight run, a BLOCKED report that you skim at 8am reads as "no findings = clean"** — and the bug it never looked at ships to the client. This is exactly the failure class this repo exists to prevent ("factual errors ship because a s