ICD疾病分类与手术操作编码技能。熟练掌握中国ICD-10、ICD-9-CM-3、ICD-O编码体系，精通主要诊断选择原则（含黄锋版规则），熟悉DRG/DIP分组规则，支持诊断/手术主导词双向查找（内置词条数据库）。

Do not run this skill against real patient data until you verify its contents and behavior. Specific steps to take before installing or executing: 1) Confirm the data files: SKILL.md lists large datasets (data/clinical_terms.json, data/drg_weight.json, data/drg_icd_map.json). Check whether those files are actually included in the package — if they're missing, the included Python tools may attempt to download data or contact remote servers. 2) Inspect the Python scripts (docs/icd_lookup_tool.py and docs/drg_lookup_tool.py) before execution: look for network calls (requests, urllib, socket), subprocess/OS calls, file writes outside the package, or code that collects environment variables. If you cannot read them yourself, ask someone with Python experience to do so. 3) If scripts perform network access, verify the endpoints are legitimate (WHO, national code sites) and that data sent is limited to queries — do not allow exfiltration of PHI/secret files. Prefer running the scripts in a sandbox or offline environment initially. 4) Source validation: the skill's homepage and author source are unknown. Prefer packages with clear provenance for clinical tools; verify licensing and whether the data mapping is permitted for your use (regulatory/privacy considerations). 5) Operational safety: because this tool influences coding and potentially billing (DRG/DIP), review outputs with an experienced coder/clinician before relying on them for official records. If any Python script makes unexpected network calls, attempts to read unrelated local files (e.g., /etc/, home directories) or writes back to external endpoints, treat the skill as unsafe and do not enable it in production.

Type: OpenClaw Skill Name: icd-coding Version: 2.2.0 The skill bundle is a highly detailed medical coding assistant focused on ICD-10, ICD-9-CM-3, and CHS-DRG 2.0 systems. It contains Python utility scripts (icd_lookup_tool.py and drg_lookup_tool.py) designed to parse medical Excel files and perform lookups against local JSON databases. The documentation is purely educational and referential, providing extensive medical knowledge and coding rules. No evidence of malicious intent, data exfiltration, unauthorized network access, or harmful prompt injection was found.