菜单
Review a spec, proposal, or design before implementation to catch gaps, risks, and ambiguities. Runs multi-angle parallel review by default; scans codebase for context first. Use --fix for iterative review+fix loop (agent escalates design-judgment issues to you), --angles to narrow the angle set.
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
Inject reviewer-aligned rules into openspec/config.yaml so generated specs pass review on first attempt. Use to reduce review round-trips after /opsx:ff or spectra-propose.
Review implementation against a spec using git diffs to catch mismatches, missing work, and bugs. Supports iterative review+fix with --fix flag and parallel multi-angle review. Use after implementing a feature to verify spec coverage.
Dual-engine spec review — Claude and Codex review the same change in parallel, then a Workflow loops fixes until BOTH engines are MEDIUM-clean. Use when a spec is high-stakes and you want two different models to cross-check for blind spots before implementation. Requires the openai-codex plugin for the Codex engine.
Frame a feature, bug, or design question before building: scout the codebase, surface assumptions or interview, compare options, and converge on an explicit decision. Use when an idea needs sharpening before a spec or plan exists. Read-only — never implements.
Synthesize git activity across multiple projects over a time window into themed recall notes. Clusters commits into themes (features, fixes, refactors, infra, learning surface, tickets, side quests), auto-hides empty sections, and surfaces work that PM tools don't track. Use when the user asks 'what did I do last week', wants a multi-project activity summary, or needs to recall their development history.
Merge git branches with auto-generated Conventional Commits messages. Use when merging branches, finishing feature work, or creating merge commits.
| name | reviewer:spec |
| description | Review a spec, proposal, or design before implementation to catch gaps, risks, and ambiguities. Runs multi-angle parallel review by default; scans codebase for context first. Use --fix for iterative review+fix loop (agent escalates design-judgment issues to you), --angles to narrow the angle set. |
Review feature specifications, proposals, designs, or task lists before implementation starts.
Every invocation fans out across all built-in review angles by default (bound engineering — first pass extracts maximum signal). Flags only narrow that set or layer a fix loop on top.
Provide the paths to spec files or folders to review:
$spec docs/plans/auth-flow/
$spec proposal.md spec.md tasks.md
With iterative fix loop (agent decides which issues need your call):
$spec docs/plans/auth-flow/ --fix
$spec docs/plan/ --fix -n 5
Narrow to specific angles (still parallel, just fewer):
$spec docs/plan/ --angles "scope,tasks"
Skip codebase exploration:
$spec docs/plan/ --no-explore
If no paths are given, ask which files to review.
--no-explore).claude/reviewer/rules.yaml at the git root if it exists (harvested project rules); pass its content as project_rules (empty if absent)fix_enabled derived from --fixBefore dispatching the orchestrator, the code-explorer agent scans the codebase for relevant context based on spec content. Pass --no-explore to skip this step.
Always dispatch reviewer:spec-orchestrator with parameters (paths, fix_enabled, max_iterations, angles, codebase_context, review-angles template content, log_script_path (resolve ${CLAUDE_PLUGIN_ROOT}/scripts/log-findings.sh to an absolute path), project_rules). The orchestrator fans out per angle in parallel and decides whether to also run the composition angle based on whether the scope contains multiple independent spec units. When fix_enabled is true it also runs Steps 5–7, escalating only the issues that need design judgment.
The agent will cd to the git root automatically. Provide it with: