Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

A Share Model Risk · aifinlab | DecimalAI

financeby aifinlab

A Share Model Risk

aifinlab/a-share-model-risk

Source aifinlab/finclaw ↗License Apache-2.0

A股模型风险/回测过拟合分析。当用户说"模型风险"、"model risk"、"过拟合"、"overfitting"、"回测失真"、"样本外失效"、"模型验证"时触发。基于 cn-stock-data 获取数据，评估量化模型的过拟合与模型风险。支持 formal/brief 两种输出风格。

SkillScore87Maintained

Automatic

Finance & Accounting pack

Last verified 8/27/2026

What it changes

Passed 20 of 23 test cases with the skill, and 19 of 23 without it.

Without skill83%

With skill87%

Checks it passed that the model alone missed

The output starts with header `## 模型风险速览`.

1 fixed19 passed either way3 still failing0 broke

Pass-rate lift: +4 pp
Tokens: +17%
Agent turns: 0%
Response time: +6%

Model gemini-3.6-flashMeasured 8/27/2026v120 of 23 cases passed with skill

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/aifinlab-a-share-model-risk/SKILL.md https://app.decimal.ai/s/aifinlab-a-share-model-risk/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add aifinlab-a-share-model-risk, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/aifinlab-a-share-model-risk/ and pins it in .decimal/skills.lock

decimalai skills add aifinlab-a-share-model-risk

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/aifinlab-a-share-model-risk/SKILL.md https://app.decimal.ai/s/aifinlab-a-share-model-risk/SKILL.md

Raw file: /s/aifinlab-a-share-model-risk/SKILL.md ↗ · machine-readable summary at /s/aifinlab-a-share-model-risk.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

A test case the skill fixed. Same model, same prompt, with and without the skill.

Prompt

对某 A 股小市值策略进行模型风险快速诊断，训练 Sharpe 3.5，测试 Sharpe 1.8（落差 48.5%）。请按简报格式输出分析结果。

What was checked, without and with the skill
Checked	Without skill	With skill
The output starts with header `## 模型风险速览`.	Missed	Passed
The output explicitly states the performance gap percentage.	Passed	Passed

case-17 · gemini-3.6-flash · 8/27/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md66 lines2 files

模型风险/回测过拟合分析助手

数据获取

通过 cn-stock-data skill 获取数据：

回测结果: 策略回测的绩效数据
模型参数: 模型配置与超参数
样本外数据: 未参与训练的数据

分析工作流

Step 1: 过拟合检测

训练集vs测试集表现差距：差距>30%疑似过拟合
参数敏感性：微调参数后表现大幅变化=过拟合
策略复杂度：参数越多过拟合风险越大
多重检验：测试了N个策略，需调整显著性水平

Step 2: 回测陷阱识别

前视偏差(look-ahead bias)：使用了未来信息
幸存者偏差(survivorship bias)：只用存活股票
数据窥探(data snooping)：反复优化直到好看
交易成本低估：未考虑滑点/冲击/流动性

Step 3: 模型验证方法

样本外测试：严格的时间外推验证
Combinatorial Purged CV：Lopez de Prado方法
Walk-forward分析：滚动窗口前推验证
蒙特卡洛排列检验：随机打乱标签的基准

Step 4: 模型风险管理

模型多样化：不依赖单一模型
定期重验证：每季度重新评估模型有效性
模型衰减监控：IC/Sharpe的滚动趋势
模型退役标准：连续N月表现低于阈值则停用

Step 5: 输出报告

输出格式

formal 风格（研报级）

Track record

Benchmark

20 / 23

test cases passed, vs 19 / 23 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 v4 — 8/3/2026

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

Imported

aifinlab / finclaw

skills/a-share-model-risk/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/3/2026Version v1Bundled 1 file

More skills from the same source and nearby in the registry, with their headline benchmark lift.