FinTech Risk Control Expert
v1.0.0数字金融科技与风控策略专家。当用户要求进行数据分析、使用Python处理金融数据、构建风控模型（决策树、分箱）、进行特征工程与分箱、分析信用风险、生成风控规则、构建评分卡等场景时使用此技能。
⭐ 0· 91·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (financial risk modeling, WOE/IV, decision trees, scoring) match the SKILL.md content. The skill's examples and functions (pandas, numpy, sklearn) are exactly what you'd expect for the described capabilities; there are no unrelated credentials, binaries, or config requirements.
Instruction Scope
SKILL.md is an instruction-only document with Python snippets that operate on local CSV input, compute WOE/IV, build decision trees, and export rules. The instructions reference only data processing and ML tasks relevant to the stated purpose. Minor coding issues (e.g., numeric_cols is referenced but not defined) are implementation bugs rather than scope creep.
Install Mechanism
No install spec and no code files; nothing is downloaded or installed by the skill. This minimizes persistence and external install risk. The runtime does assume a Python environment with pandas/numpy/sklearn available, but the skill does not attempt to install them.
Credentials
The skill declares no environment variables, no credentials, and no config paths. That is proportionate for an instruction-only ML/risk-modeling helper that works on local data files.
Persistence & Privilege
always is false and model invocation is not disabled (normal). The skill does not request permanent presence or system-level changes and does not modify other skills or system configuration.
Assessment
This skill is instruction-only and coherent with its stated purpose, but before using it: (1) ensure the agent's runtime has the expected Python libraries (pandas, numpy, scikit-learn); (2) never feed sensitive production data unless you trust the execution environment — test on anonymized or sample data first; (3) review and/or run the provided code snippets in a sandbox because there are minor bugs (e.g., undefined numeric_cols) and you should validate that outputs meet your regulatory and business requirements; (4) if you expect the agent to run these snippets autonomously, confirm the agent does not have network or filesystem permissions you don't intend to allow.Like a lobster shell, security has layers — review code before you run it.
credit-scoringvk970729vqcgbgcy7nzeke2aj6x83j6gqfintechvk970729vqcgbgcy7nzeke2aj6x83j6gqlatestvk970729vqcgbgcy7nzeke2aj6x83j6gqrisk-controlvk970729vqcgbgcy7nzeke2aj6x83j6gq
License
MIT-0
Free to use, modify, and redistribute. No attribution required.