AI 应用工程师日报
v1.0.0AI 应用工程师专属的每日资讯日报技能。收集 AI 行业动态（AI 龙头公司官方博客/权威来源），按六大板块组织内容，面向 AI 应用开发工程师，重点关注能力落地和技术实用。当用户说"AI 日报"、"每天 AI 资讯"、"帮我看看今天 AI 圈发生了什么"、"生成一份日报"时触发。
⭐ 0· 70·0 current·0 all-time
byAriel Overcast@robusth
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 (daily AI news for application engineers) matches the declared behavior: SKILL.md lists public AI blogs, media, and community sources and the expected 6+1 section output. Nothing requested (no env vars, no binaries) is disproportionate to the stated purpose.
Instruction Scope
Instructions direct the agent to '抓取' (fetch) specified public sites, filter arXiv titles, and produce a structured markdown report including source URLs. This is exactly what the skill needs to do, but the SKILL.md is implementation-agnostic about how to fetch (e.g., web request tool vs. browser). That grants the agent broad discretion to access the web during invocation — expected for this skill, but operators should ensure the agent's web-access tooling is appropriately constrained and conforms to robots/rate limits.
Install Mechanism
No install spec and no code files are included (instruction-only). Nothing will be written to disk or downloaded by the skill itself during install, which minimizes install-time risk.
Credentials
The skill declares no required environment variables, no primary credential, and no config paths. SKILL.md does not request any secrets or access to unrelated services. The lack of credential requests is proportionate to a news-aggregation skill.
Persistence & Privilege
always is false and the skill does not request persistent privileges or modify other skills or system settings. Autonomous invocation is allowed (platform default) but not combined here with elevated access or credentials.
Assessment
This skill is an instruction-only news aggregator that fetches public AI blogs and media and produces a Chinese daily briefing; it asks for no credentials and installs nothing. Before enabling: (1) confirm you are comfortable with the agent having web-access to fetch the listed sites (ensure tooling enforces allowed hosts and rate limits), (2) be aware it will include source URLs in outputs so verify any sensitive links yourself, (3) if you need different sources or stricter filtering (e.g., exclude specific domains), edit the SKILL.md or restrict the agent's network capabilities, and (4) remember the static scanner had no files to analyze — absence of findings only reflects that this is an instruction-only skill, not a guarantee of safety.Like a lobster shell, security has layers — review code before you run it.
latestvk97cp4ez42n6e3wd123gs5z4r9841p2n
License
MIT-0
Free to use, modify, and redistribute. No attribution required.