forecast-analysis-claw
v1.0.0根据历史销售数据预测未来销量并生成补货建议。核心能力：(1) 销量预测 - 基于移动平均、指数平滑、Holt-Winters、Prophet 等模型自动预测未来销量；(2) 补货计算 - 结合库存参数自动计算补货触发点和建议补货量；(3) 活动预测 - 叠加促销效应系数预测大促期间销量峰值；(4) 断货预警 -...
⭐ 0· 50·0 current·0 all-time
byRicky@tujinsama
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name/description (sales forecasting & replenishment) align with included scripts and reference docs. Required inputs are CSV/XLS data and optional inventory params — no unrelated credentials, binaries, or config paths are requested.
Instruction Scope
SKILL.md instructs running local Python scripts on user-provided files (clean, single, batch, report, evaluate). The instructions reference only local files and the included scripts; they do not instruct reading system config, environment secrets, or transmitting data externally.
Install Mechanism
No automated install spec; dependencies are standard Python packages (pandas, numpy, scikit-learn) and optional prophet. pip installs are expected for this kind of tool and are proportionate to functionality.
Credentials
The skill requests no environment variables, credentials, or config paths. The code only reads user-supplied data files. No secret-exposing env var usage detected.
Persistence & Privilege
Skill is user-invocable only (always:false). It does not request permanent presence or modify other skills or system-wide settings.
Assessment
This skill appears coherent for forecasting and replenishment: it runs local Python scripts on your CSV/XLS sales data and doesn't ask for credentials or network endpoints. Before installing/using it: (1) run it in an isolated environment (virtualenv/container) and review the full forecast-runner.py file yourself (the provided snippet was truncated in the review); (2) be aware you must pip-install standard data science packages (pandas, numpy, scikit-learn) and optionally prophet/statsmodels; (3) the scripts process your raw sales data locally and do not include obvious exfiltration, but you should still test with non-sensitive sample data first; (4) check edge cases (divisions by zero in evaluation when actual=0) and validate outputs before automating orders based on recommendations.Like a lobster shell, security has layers — review code before you run it.
latestvk975cgscrfdy80jteh3knr0nqx84fjxh
License
MIT-0
Free to use, modify, and redistribute. No attribution required.