Agent Evaluation Report
v1.0.0根据测试数据自动生成标准化的智能体系统评测报告。 Use when: 用户说"生成智能体评测报告"、"创建测试报告"、"项目测试报告"。
⭐ 0· 47·0 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
Name/description match the SKILL.md: it generates evaluation reports from test data. Declared capabilities (Feishu doc operations, read/write, message) are consistent with producing and saving reports.
Instruction Scope
Instructions stay within report generation (template, required inputs, save locations). One minor note: SKILL.md includes a hard-coded Feishu knowledge-base ID and node (7616288931050507220 / 效果评测/测试报告), which means the skill will attempt to write to a specific target if Feishu access is available — this is consistent with a doc-writing skill but worth confirming you want that target.
Install Mechanism
No install spec and no code files beyond SKILL.md/package.json. Instruction-only skills carry low install risk because nothing is downloaded or executed on install.
Credentials
Skill declares no required environment variables or credentials, which is reasonable. It uses Feishu tools (feishu_create_doc etc.); platform-level Feishu credentials/permissions (not declared in the skill) will be needed at runtime — confirm that Feishu access granted to the agent is appropriate for writing to the specified knowledge base.
Persistence & Privilege
always is false and the skill does not request elevated persistence. It writes output to a local folder (output/effect-reports/) and to Feishu docs per its instructions — this is expected behavior for a report generator.
Assessment
This skill appears to do what it says: generate standardized agent test reports and save them as Markdown/Word and to Feishu. Before installing, confirm: (1) you trust the skill source (owner unknown) and are comfortable with the agent writing to the hard-coded Feishu knowledge base/node listed in SKILL.md; (2) the platform's Feishu credentials granted to the agent have the intended scope (so it can't write to unexpected org resources); (3) any sensitive test data you provide will be written to output/effect-reports/ and to Feishu — review outputs before sharing. No environment variables or external installers are required by the skill itself.Like a lobster shell, security has layers — review code before you run it.
latestvk97ef4gjwsgk668wnwvv5eakkx84r09m
License
MIT-0
Free to use, modify, and redistribute. No attribution required.