Ai Intel Radar Repo

Build a high-signal AI intelligence brief from fresh web sources, with separate daily and weekly formats. Use when the user asks for latest AI news, trend mo...

Security Scan

The skill's requirements and runtime instructions are consistent with its stated purpose of producing AI news briefs; it is instruction-only, asks for no credentials, and references reasonable public sources.

Purpose & Capability

Name, description, and listed sources align: the skill's goal is to build AI intelligence briefs and the SKILL.md plus references/sources.md focus on public AI news, papers, and blogs. There are no unexpected credentials, binaries, or config paths requested.

Instruction Scope

Instructions are scoped to collecting recent public AI reporting and producing strict daily/weekly formats in Chinese by default. They do not instruct reading local files or environment secrets. However, the skill leaves the fetching mechanism and exact allowed endpoints somewhat open (it points to a tiered source list but doesn't hard-limit the agent), which grants discretion in web requests and source selection.

Install Mechanism

No install spec and no code files — instruction-only. No downloads, packages, or binaries are installed, so there is minimal install-time risk.

Credentials

The skill requests no environment variables, credentials, or config paths. The safety note in SKILL.md explicitly calls out not exposing secrets. Required access (web browsing) is proportionate to the stated purpose.

Persistence & Privilege

always is false and the skill does not request persistent system-level privileges or modify other skills. It can be invoked by the agent (platform default), which is appropriate for this utility.

Assessment

This skill appears coherent and safe: it only directs the agent to summarize public AI sources and asks for no credentials. Before enabling it, consider: (1) the agent will fetch web content — if you have strict browsing/privacy constraints, confirm that web access is permitted; (2) the SKILL.md grants the agent discretion to choose sources beyond the listed tiers, so verify outputs (links and claims) before acting on them; and (3) if you need output in a different language or to exclude paywalled/specific regional sources, ask for that explicitly in prompts.

AI Intel Radar

Produce concise, decision-ready AI intelligence in Chinese by default.

Mode selection

• If user asks "today" / "daily" / "最新": use Daily mode
• If user asks "this week" / "weekly" / "本周": use Weekly mode
• If unspecified: default to Daily mode

See source guidance in references/sources.md.

Daily mode output (strict)

1. 3条快讯（发生了什么）
2. 2条技术深度（为什么重要）
3. 1条商业判断（潜在影响）
4. 3条可执行建议（本周可做）

Each item must include:

• 结论一句话
• 1条来源链接
• 时间（若可得）

Weekly mode output (strict)

1. 趋势变化（3条）
2. 机会清单（3条）
3. 风险清单（3条）
4. 本周Top 3行动

Quality rules

• Prefer official and primary sources first.
• Avoid hype and repeated rumor chains.
• Mark uncertain claims as "待验证".
• Do not dump long lists; keep high signal only.
• If source freshness is weak, explicitly say so.

Safety

• Never expose secrets/tokens.
• Never fabricate links.
• If no reliable updates are found, return a short "低变化日" summary.