iamzifei/aikey / skills/aikey-review
aikey-review
VerifiedAI KEY·发布前审一遍 ——口播稿发布前审核技能。按观众的四次决定审:点不点进来 · 留不留下来 · 记不记得你 · 做不做点什么。先盲读开头判留走,逐句信息密度评分(60/80 分线)+ 十一问 + 六种反应预测(只算观众走得到的部分,含流量是谁、离生意多远)+ 红线六查(改法给稳妥版和保留力度版两版)+ 机器信号层(导流 / 广告形状 / 名单词,与内容违规分开报),默认只诊断不改。不负责重写整稿(→ aikey-script)、换开头或标题(→ aikey-hook / aikey-title)、拆别人的爆款(→ aikey-resonate)。 触发方式:/aikey-review、/能不能发、/审核、/aikey-审核、「这稿子能不能发」「帮我审一下」「过一遍红线」「信息密度够不够」 Pre-publish review for talking-head scripts, organised around the viewer's four decisions: click, stay, remember, act. Per-sentence density scoring, eleven questions, red-line audit. Diagnose-only by default. Not for: rewriting the script (aikey-script), replacing the opening or title (aikey-hook / aikey-title), decoding someone else's hit (aikey-resonate). Trigger: /aikey-review, "can I publish this", "review my script" —— AI KEY · 不给公式,给判据。每条规则都标了实测代价。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 29.5 KB |
| evals/README.md | markdown | 4.8 KB |
| evals/sample-01-导流与卖真.md | markdown | 1.3 KB |
| evals/sample-01-答案.md | markdown | 1.4 KB |
| evals/sample-02-开头空好料在后.md | markdown | 1.2 KB |
| evals/sample-02-答案.md | markdown | 1.4 KB |
| evals/sample-03-开头硬但好反驳.md | markdown | 859 B |
| evals/sample-03-答案.md | markdown | 1.3 KB |
| evals/sample-04-开头误会题目.md | markdown | 1.1 KB |
| evals/sample-04-答案.md | markdown | 1.7 KB |
| evals/sample-05-答案.md | markdown | 1.5 KB |
| evals/sample-05-语境红线.md | markdown | 874 B |
| evals/sample-06-技能该管的地方.md | markdown | 746 B |
| evals/sample-06-答案.md | markdown | 1.3 KB |
| references/反应预测.md | markdown | 8.4 KB |
| references/规则卡-速查.md | markdown | 13.5 KB |
| references/规则卡.md | markdown | 27.8 KB |
| references/评分锚点.md | markdown | 7.2 KB |
evals/sample-02-答案.md
View on GitHub# 测试稿 02 · 答案(只给评分者)
考的是 v0.5.0 新加的「开头先盲判」。第一句是空话,全稿的货(四百封询盘、一小时内回复成单率是隔天的六倍)排在第七段。
## 必须抓到(5)
| # | 要求 |
|---|---|
| B1 | 报告里有**只基于第一句**的判断(「他以为要讲什么 / 多少人心里有 / 能拿到什么」或等价三件),且这段判断里**没有引用第二句以后的内容** |
| B2 | 开头判定为「走」或「犹豫」(不是「留」);理由落在第一句空、没新东西、听过很多遍这一类 |
| B3 | 明确点出「一小时内回复成单率是隔天六倍」(或四百封询盘分组那段)是全稿最强的一处,且**大多数人听不到** / 排在观众划走之后 |
| B4 | 改法是把这处挪到开头(或以它重写第一句),**不是另造一句夸张的钩子** |
| B5 | 第 2 问的结论没有被后文翻案(没有出现「开头弱但后面很好所以整体可以留住人」这种抵消式表述) |
## 不许误报(扣分)
| # | 误报 |
|---|---|
| X1 | 因第七段数据好而把开头判为「留」或「合格」 |
| X2 | 把「六倍」「四百多封」当卖真红线违规(这是过程数据,不是收入金额) |
## 评分
B1–B5 各 1 分,X1/X2 各扣 1 分。满分 5,通过线 4 且 X1 未命中。
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |