仅适用于北京市企业职工基本养老保险退休金测算。支持依据北京政策参数计算基础养老金、个人账户养老金、过渡性养老金，并对未退休用户做未来缴费策略优化。

仅适用于北京市企业职工基本养老保险退休金测算。支持依据北京政策参数计算基础养老金、个人账户养老金、过渡性养老金，并对未退休用户做未来缴费策略优化。

What to check before installing/using: - This skill runs local Python scripts and (optionally) tesseract OCR on files you upload (CSV/XLSX/images). It processes personally sensitive data (birth dates, contribution history, account balances), so run it in a trusted/local environment and avoid sending inputs to unknown remote services. - There are no network calls or credential requirements in the code; references to official policy URLs are documentation only. - You must install Python >=3.10 and dependencies (pandas, openpyxl) and tesseract+chi_sim if you plan to use OCR. - The code contains some minor robustness bugs (e.g., a likely typo in the OCR parsing function and a truncated variable in calc script seen in provided files) that may cause runtime exceptions; review or test the scripts before relying on results. - If you do not want autonomous invocation by an agent, restrict the skill’s invocation policy; by default model invocation is allowed but always:false here means it is not force-included. - Recommended: inspect the scripts yourself (or run in an isolated environment), validate outputs against known examples, and verify that parameter updates (scripts/update_params.py) are only performed intentionally.

Type: OpenClaw Skill Name: retirecalc-beijing Version: 1.0.1 The skill bundle is a legitimate and highly specialized tool for calculating Beijing enterprise employee pensions according to local regulations (e.g., Beijing Decree No. 183 and the 4050 subsidy policy). It includes a robust pipeline for data ingestion from CSV, XLSX, and images using Tesseract OCR (scripts/ingest_user_data.py), and a detailed calculation engine (scripts/calc_pension_beijing.py). The code follows standard practices, uses list-based subprocess calls for safety, and contains extensive documentation and policy references (references/policy-links.md) consistent with its stated purpose. No evidence of malicious intent, data exfiltration, or harmful prompt injection was found.