菜单
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
Run live continuous co-review or replay persisted reviewer evidence for an iteration.
Perform a Specrew human-verdict boundary stop on the Claude host. Renders the FULL Rule 46 six-section human re-entry packet AND the verdict options as one Markdown message, with the AskUserQuestion picker disabled so the packet cannot collapse into the picker's short header/option fields. Invoke at EVERY human-judgment boundary stop (specify, clarify, plan, tasks, before-implement, implement, review, retro, feature-closeout, lifecycle-end). Triggers: boundary stop, verdict, approve / redirect / send back, why I stopped, human re-entry packet, gate stop.
{what this skill teaches agents}
Run Specrew's per-lens design workshop and collaborative design-analysis. Use whenever you work the design lenses for a feature: at specify/intake (the lens workshop) and at the design-analysis stop (co-design the architecture), and RE-INVOKE at the start of EACH new lens (architecture, data, ui-ux, security, integration, devops, requirements/NFR, observability, component). Triggers: design, design lens, lens workshop, design-analysis, architecture, trade-offs, co-design, explore options, decompose, or moving from one lens to the next. Tells you to facilitate each lens as a discussion, surface diagrams the human can actually SEE (console ASCII inline; mermaid/html to a file with a clickable file:/// link), co-design components/responsibilities/flows WITH the human instead of handing over finished options, capture the agreements, and which per-lens md to load.
Re-load scoped Specrew methodology discipline into context on demand — the manual recovery surface for agent drift, and the engine behind the automatic refocus injections. Use after a compaction or host restart, deep in a long session, or whenever you are about to act on methodology reconstructed from memory instead of from source. Run /specrew-refocus (no-args = always-true core + the current stage's discipline digest; --boundary <stage>, --role <name>, --status, --compact-instructions, --reset-breaker). Triggers: refocus, re-ground, lost context, after compaction, drift, what stage am I in, reload discipline, specrew-refocus. Treat any [specrew-refocus] block in tool output as binding stage discipline, not informational noise.
| name | specrew-review |
| description | Run live continuous co-review or replay persisted reviewer evidence for an iteration. |
Type: Operational Skill
Schema: v1
Status: Active
Namespace: /specrew
Canonical command: /specrew-review
Trigger or inspect the Specrew review-oriented workflow. Use live mode to run the continuous co-review runtime and persist reviewer evidence under .specrew/review/inline/<run-id>/; use replay mode to inspect an existing reviewer closeout packet for a completed iteration.
This skill supports review work but does not bypass the required human decision point. A human reviewer must still explicitly approve lifecycle advancement. Invoking /specrew-review is not equivalent to approving an iteration for closure or advancement.
Feature 016 rule: Humans remain the approval authority for major boundary transitions. This command may not imply authorization to advance from any lifecycle boundary.
This skill is part of the /specrew-* command surface. It coexists with /speckit.* review-related commands without collision or shadow. Using /specrew-review does not substitute for /speckit.* lifecycle actions that require explicit human authorization.
/specrew-review [<iteration>] [--project-path <path>] [--feature <id>]
[--iteration <NNN>] [--quiet] [--json] [--open]
/specrew-review --live [--baseline-ref <git-ref>] [--project-path <path>] # baseline auto-anchors for signoff; an explicit --baseline-ref run is exploratory-only
[--checkpoint-id <id>] [--run-id <id>] [--host <host>]
[--model <model>] [--effort <effort>] [--authorization-ref <ref>]
[--code-writer-host <host>]
[--design-context-ref <path>] [--allowed-path <path>]
[--forbidden-path <path>] [--exclude-path <path>]
[--reviewer-config <path>] [--schema-root <path>]
[--run-root <path>] [--timeout-seconds <N>]
[--fallback-policy <none|one-authorized-availability-fallback>]
[--preserve-debug]
Backed by: specrew review / scripts/specrew-review.ps1
| Input | Type | Required | Description |
|---|---|---|---|
<iteration> | string (positional) | No | Iteration number to replay (defaults to most recent closed iteration) |
--project-path | string | No | Target Specrew project path (defaults to current directory) |
--feature | string | No | Restrict lookup to one feature directory under specs\ |
--iteration | string | No | Replay a specific iteration directory |
--quiet | flag | No | Emit only the stable machine-parseable digest line |
--json | flag | No | Emit JSON summary instead of the visual reviewer summary |
--open | flag | No | Open reviewer-index.md and review-diagrams.md when present |
--live | flag | No | Run the continuous co-review runtime instead of replaying prior closeout state |
--baseline-ref | string | Yes for --live | Git ref used to compute the review change set |
--checkpoint-id | string | No | Stable checkpoint identifier for the live review request |
--run-id | string | No | Explicit run id for evidence under .specrew/review/inline/<run-id>/ |
--host | string | No | Reviewer host to request (claude, codex, copilot, cursor-agent, or antigravity) |
--model | string | No | Reviewer model override; non-default models require --authorization-ref |
--effort | string | No | Optional host-specific reviewer reasoning/effort label persisted in evidence |
--authorization-ref | string | No | Human authorization reference for host/model selection |
--code-writer-host | string | No | Implementation host (claude, codex, etc.) used to prefer an independent peer reviewer |
--design-context-ref | string | No | Design or lifecycle artifact to include in the review request context |
--allowed-path | string | No | Add a mutable path allowlist entry for this live review run |
--forbidden-path | string | No | Add a protected path entry for this live review run |
--exclude-path | string | No | Add a change-set exclusion entry for this live review run |
--reviewer-config | string | No | JSON reviewer host configuration for deterministic or authorized hosts |
--schema-root | string | No | Contract schema root for ReviewRequest and FindingsResult validation |
--run-root | string | No | Override the evidence output root; defaults to .specrew/review/inline |
--timeout-seconds | number | No | Live reviewer host timeout |
--fallback-policy | string | No | Host fallback policy (none or one-authorized-availability-fallback) |
--preserve-debug | flag | No | Preserve debug artifacts from the live review run |
ReviewRequest, prompt bundle, host invocation metadata, FindingsResult, gate verdict, and run index under .specrew/review/inline/<run-id>/ when --live is used.specrew review workflow with minimal wrapper context.Only the arguments listed above are accepted in v1. Unknown arguments are rejected with command-specific help guidance.
| Failure mode | Behavior |
|---|---|
| Unsupported argument | Rejected immediately with command-specific help guidance |
Missing --baseline-ref in live mode | Error explaining that live review needs a baseline ref |
| Unauthorized reviewer host/model | Infrastructure failure with reviewer authorization guidance; do not treat this as reviewer approval |
| No iteration found | Error with guidance to check specs/ directory structure |
| Missing project setup | Stop with specrew init remediation |
/specrew-where — current project status (live, not review replay)/specrew-help — catalog fallback and full command list