원클릭으로
apply-review
Load saved review feedback from /bpe:review and apply changes to the reviewed artifact
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Load saved review feedback from /bpe:review and apply changes to the reviewed artifact
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Iterative Q&A to develop a thorough project specification (spec.md)
Generate a commit message explaining what was changed and write it to commit-msg.md
Implement the next unchecked step from plan.md by executing its sub-steps as written (RED-GREEN-REFACTOR for Feature steps; Scope/Tooling/Do/Verify/Document for Task steps)
Fetch a GitHub issue and route to brainstorm or plan based on detail level
Autonomous-mode BPE run via /goal. Modes — full (default) | section <name> | step. Pre-flights branch safety (refuses on main), resolves the project verification command (test-runner autodetect, spec.md fallback, or ask), builds a verifiable completion condition, and writes the assembled /goal block (condition + validator-aware orchestrator playbook + per-commit verification) to goal.md at the repo root for you to paste. Requires Claude Code v2.1.139+; put your session in auto mode before pasting for unattended execution.
Create, continue, or close handoff documents for fresh-agent baton passes
SOC 직업 분류 기준
| name | apply-review |
| description | Load saved review feedback from /bpe:review and apply changes to the reviewed artifact |
| argument-hint | [feedback-json-path] (optional; defaults to the most recent /tmp/bpe-review-*-feedback.json) |
| model | opus |
| disable-model-invocation | true |
Consume the feedback JSON written by /bpe:review and apply the user's annotations to the reviewed markdown artifact. The user has already reviewed in the browser and clicked Save — this command surfaces their decisions and applies the changes after explicit confirmation.
If the user passed a feedback path as an argument, use that file. This is the escape hatch for replaying an older review or for the case where the user ran two reviews in parallel and the wrong one is "most recent."
Otherwise, find the most recent feedback JSON in /tmp/:
ls -t /tmp/bpe-review-*-feedback.json 2>/dev/null | head -1
If none exists, tell the user to run /bpe:review first and stop.
If the chosen file's saved_at timestamp is more than 24 hours old, surface that to the user and ask whether they meant to apply this older feedback or run /bpe:review again.
Read the feedback JSON. Expected shape:
{
"sections": [
{
"id": "section-N",
"heading": "<heading text>",
"decision": "ship" | "update" | "redirect" | "reject" | "unset",
"comment": "<free-text feedback>"
}
],
"global_comment": "<free-text>",
"artifact_path": "<absolute path to the reviewed markdown file>",
"saved_at": "<ISO timestamp>"
}
Each entry is one decision unit — the review is intentionally fine-grained, so expect many entries. Display a summary that front-loads the destructive decisions, since those are what the user most needs to verify before saying yes in Step 3:
N reject, N redirect, N update, N unset, N ship.ship entries under sub-headings, in this exact order (most-destructive first):
reject) — list each unit's heading + comment. These are about to be removed; surface them first so the user can catch a mistake before authorizing.redirect) — list each unit's heading + comment. These are full rewrites driven by the comment.update) — list each unit's heading + comment. These are tweaks.unset) — list any units that have a comment (those'll be treated as update) and any that don't (those'll be left alone).ship units don't need to appear in the summary — they're the silent majority by design and listing them adds noise. One exception: a ship entry with a non-empty comment. The review page locks the comment box while Ship it is selected, so this only appears in feedback saved by an older page — but when it does, list it under a Ship, with comment sub-heading rather than dropping the comment silently.
Ask the user explicitly: "Apply these changes to <artifact_path>?" Do not proceed without a clear yes.
How to interpret each decision (increasing severity):
update directive or ignore it — never discard it without surfacing it.Apply the global_comment as a holistic directive across the artifact (e.g. "tighten language", "add a glossary", "split into two files").
Read the artifact via the Read tool. Make targeted edits via the Edit tool — one Edit per affected section. Do not rewrite the whole file unless the global comment explicitly requests it. Preserve formatting, list styles, and unrelated sections exactly.
After applying, show the user:
/bpe:review again to re-review, /bpe:plan to regenerate downstream artifacts (if spec.md was edited), or continue.