code reviewby ontology-of-everything
Jackson 概念审计
ontology-of-everything/jackson-concept-audit
Audits an existing codebase against its Jackson concept model (spec drift, boundaries, criteria, composition, dependencies).
What it changes
Passed 16 of 21 test cases with the skill, and 3 of 21 without it.
Without skill16%
With skill76%
Checks it passed that the model alone missed
- Identifies user control takeover as over-sync.
- Routes naming finding to `jackson-concept-design`.
- Identifies missing automated trigger as under-sync.
and 17 more in the full report
15 fixed4 passed either way6 still failing0 broke
- Pass-rate lift
- +60 pp
- Tokens
- +2%
- Agent turns
- 0%
- Response time
- -34%
Model gemini-3.6-flashMeasured 9/27/2026v119 of 25 cases passed with skill · 21 comparable
See it in action
One of 15 test cases the skill fixed. Same model, same prompt, with and without the skill.
Prompt审计 `UserAccountManager` 模块时，发现其同时承担用户身份认证与信用卡积分结算两项无关职责。在判据重审维度中这叫什么问题？应路由到哪里？
What was checked, without and with the skill| Checked | Without skill | With skill |
|---|
| Identifies multi-purpose module issue as conflation. | Passed | Passed |
|---|
| Routes conflation finding to `jackson-concept-design`. | Missed | Passed |
|---|
What's inside
The instructions your model reads when the skill loads.
SKILL.md51 linessingle file
Jackson 概念审计
输入：PRD 文档族（默认仓库 docs 下的 prd 目录）或概念模型 + 代码库。只读：不修改任何文件，产出发现清单；修复由路由到的技能执行。无规格文档时降级为纯边界与判据审计，并在报告中声明。
五个维度
执行步骤
- 定位规格：读 PRD 文档族与总体 PRD 的依赖图；同时利用工程自带素材（如 Spring Modulith
Documenter 生成的模块文档、cargo/dependency-cruiser 的依赖图输出）。
- 逐维度检查：每条发现记录维度、位置（文件/模块）、证据（可复核的代码或文档片段）、严重度、路由。
- 合并根因：同一问题跨维度出现时合并指向根因——概念互引的根因常是模型里未参数化的 intrinsic 依赖，而不是五条独立发现。
- 输出报告：给出修复顺序，上游优先——先模型（design）、再文档（prd）、后代码（implementation）。
报告格式
# 审计报告 <日期>
范围: <规格版本 / 代码版本>；跳过的维度及原因
## <维度名>
| 发现 | 位置 | 证据 | 严重度 | 路由 |
## 修复顺序
1. <根因级发现，上游优先>
完成条件
- 五维度全部执行，或明确声明跳过原因。
- 每条发现有位置与证据，可独立复核；无「疑似」空泛项。
- 每条发现有唯一路由；修复顺序按上游优先排列。
- 未修改任何文件。
Track record
Benchmark
16 / 21
test cases passed, vs 3 / 21 without
Adoption
- Users
- 0
- Times used
- 0
Safety and source details
SkillSafety
✓ 3 of 3 passed✓Code scan15 of 15 checks clean
0 critical · 0 warning · 0 info — scanner v5 — 8/30/2026
View 15 scanner checks
live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing
✓IntentSafe
LLM judge · does the skill do what it claims
✓ContentClean
LLM judge · policy & harmful-content review
Source
Importedontology-of-everything / semanticskills
skills/jackson-concept-audit/SKILL.md
License Apache-2.0Imported 8/30/2026Version v1