Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
F-AI 资讯搜索
v1.0.0基于Finloop权威数据源，提供金融资讯检索、实时行情和AI热闻查询，支持分类资讯、市场时段资讯和股票实时行情。
⭐ 1· 160·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description align with the actual HTTP APIs listed (news, AI hotlist, market quotes). The bundled installer and manifest are consistent with delivering an instruction-only skill into a project's .agents/skills directory. One minor inconsistency: registry metadata shows version 1.0.0 while the embedded manifest lists 1.1.3.
Instruction Scope
SKILL.md is detailed and prescriptive: it requires direct HTTP calls to the specified endpoints, mandates an automated flow to fetch all xcf_id entries and then fetch details for each, and forbids creating wrapper JS/TS files. It also says some endpoints 'may require Cookie authentication (sl-session)'. The cookie/auth requirement is not explained (no guidance how to obtain or supply 'sl-session'), which could push an agent to attempt access to existing cookie stores or other local credentials. The directives are strict but do not themselves show malicious actions.
Install Mechanism
There is no remote download/install URL; included lib/install.js and bin script copy the packaged .agents/skills files into a project's .agents/skills directory. The code shown only performs local filesystem operations and prompts before overwriting — no external code downloads or extraction from untrusted URLs.
Credentials
The skill declares no required environment variables or credentials, yet the instructions note that the banner/list endpoint 'may require Cookie authentication (sl-session)'. There is no declared primary credential, API key, or guidance for secure authentication. This mismatch (needing cookie/auth but not declaring credentials) reduces transparency and could lead to the agent trying to access local cookies or other auth material.
Persistence & Privilege
The package does write files into a .agents/skills directory when installed (via the included installer), which is expected for a skill. Flags like always:true are not set. The installer does not modify other skills or system-wide settings and requests user confirmation before overwriting.
What to consider before installing
This skill appears to be a coherent Finloop news/quotes integration, but exercise caution before installing. Things to check: 1) The SKILL.md points at UAT (ai-uat / papi-uat) hosts — confirm you want traffic sent there and whether production endpoints exist. 2) The banner/list endpoint may require a 'sl-session' cookie; the skill does not declare how to obtain or store that credential — do not supply cookies or secrets unless you trust the endpoint. 3) Review the included lib/install.js and bin script locally (they only copy files) before running npx; verify package.json and publisher identity (source is unknown). 4) If you run this in a sensitive environment, consider sandboxing network access, inspect network calls from the agent, and only provide credentials through explicit, minimal, documented env vars. If you need higher assurance, ask the publisher for production endpoints, auth methods, and a signed/verified release before installing.Like a lobster shell, security has layers — review code before you run it.
latestvk97dvvgtbna5g239py947ane3h835j8t
License
MIT-0
Free to use, modify, and redistribute. No attribution required.