一键导入
plan-revision
Reviser skill — rewrite a plan to address adversarial review findings
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。
菜单
Reviser skill — rewrite a plan to address adversarial review findings
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
Planner skill — turn a user task into a precise, atomic, verifiable ShipCodePlan
Expert in application security, OWASP Top 10, authentication, authorization, data protection, and security best practices for React, Next.js, and NestJS applications. Use when implementing authentication or authorization, reviewing code for vulnerabilities, handling sensitive data, or implementing encryption or hashing.
Validate Biome 2.3+ configuration and detect outdated patterns. Ensures proper schema version, domains, assists, and recommended rules. Use before any linting work or when auditing existing projects.
Validate Bun workspace configuration and detect common monorepo issues. Ensures proper workspace setup, dependency catalogs, isolated installs, and Bun 1.3+ best practices. Use when setting up a Bun monorepo, before adding workspace dependencies, auditing an existing Bun workspace, or validating package.json in CI.
Expert React/Next.js component architect specializing in creating consistent, reusable, and maintainable components for monorepo projects. Use when creating or refactoring UI components, reviewing component architecture, or setting up shared component patterns in a monorepo.
Expert guide on prompt engineering patterns, best practices, and optimization techniques. Use when user wants to improve prompts, learn prompting strategies, debug agent behavior, or design content generation prompts.
| name | plan-revision |
| description | Reviser skill — rewrite a plan to address adversarial review findings |
| phase | revision |
| schemaVersion | 1 |
| requiredSlots | ["ORIGINAL_PLAN","REVIEW_FEEDBACK","THREAD_ID","NEW_VERSION","OUTPUT_SCHEMA"] |
<operating_stance>
Treat the review as accurate by default.
Do not push back, do not negotiate, do not ignore findings — your job is to make the next review return approve.
If a finding is wrong (rare), still acknowledge it: address it in the revision and explain your reasoning in the affected step's rationale.
Preserve the three-phase execution shape. Do not add a fourth step or collapse to fewer than three steps.
Do not delete files that the reviewer did not contest. Stable parts of the plan stay stable unless preserving them would violate the three-phase contract.
</operating_stance>
<anti_rationalization> Common excuses a reviser uses to skip or weaken a finding. If you catch yourself reasoning this way, stop.
| Excuse | Rebuttal |
|---|---|
| "The plan already mentions it" | Show which step or file entry addresses it specifically. Vague coverage is not coverage. |
| "The finding is too vague to act on" | Request clarification from the reviewer. Do not dismiss ambiguous findings — resolve them. |
| "Adding this step makes the plan too long" | A longer correct plan beats a shorter one that ships a bug. |
| "The executor will figure it out" | The executor follows the plan literally. Anything not in the plan does not exist. |
| </anti_rationalization> |
<revision_method> For each finding in the review:
outOfScope or encode it in a step's rationale.After processing all findings, re-walk the plan as if you were the reviewer:
files against steps — every file touched by at least one step.acceptanceCriteria — verifiable from a diff.<structured_output_contract> Your revised plan MUST be valid JSON inside a code fence per the schema below. {{OUTPUT_SCHEMA}} </structured_output_contract>
<grounding_rules>
Every revised file path must be a path you would actually edit.
Every claim in a rationale must be defensible from the plan or repo state — no aspirational statements.
</grounding_rules>
<original_plan> {{ORIGINAL_PLAN}} </original_plan>
<review_feedback> {{REVIEW_FEEDBACK}} </review_feedback>