Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

BondInformation

v1.0.0

债券资讯与信用风险监测技能。基于 FEEDAX API 查询债券市场相关新闻资讯,支持按债券类型(城投债/公司债/企业债/金融债等)、情感倾向、时间范围筛选,提供债券违约风险监测、信用评级追踪、发行人舆情分析等专业功能。 使用场景:用户监测债券违约风险、追踪城投债舆情、分析公司债信用状况、查询特定债券新闻、评估债...

0· 46·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Purpose & Capability
The README, SKILL.md, and Python script all target FEEDAX bond-news queries and use an API key and the endpoint at 221.6.15.90:18011, so the functional purpose is coherent. However, the registry metadata declares no required environment variables or primary credential while the code and docs require FEEDAX_API_KEY — a clear mismatch (the skill will not work without a secret but the package doesn't advertise that).
!
Instruction Scope
SKILL.md instructs the agent/user to run 'cat .env', create a .env containing FEEDAX_API_KEY, and explicitly asks the user to "tell me API Key so I can remember". That request to paste a secret into conversational context goes beyond the stated data-access need and is unsafe: the skill should not instruct users to disclose secrets in chat. Aside from that, instructions and the script only perform HTTP queries to the FEEDAX endpoint and local file/env checks, which align with the stated purpose.
Install Mechanism
There is no install spec (instruction-only + Python script), so nothing is downloaded during installation — low install risk. The script depends on the 'requests' package (mentioned in README) but dependency requirements are not declared in registry metadata. No packaged installers or external downloads are used.
!
Credentials
The code expects FEEDAX_API_KEY (via env var, CLI, or config file) but the skill metadata lists no required credentials or primaryEnv. Additionally, SKILL.md's prompt to paste the API key into chat is disproportionate and may lead to secret disclosure. The network endpoint is a raw IP (221.6.15.90) rather than a documented domain, which reduces transparency about where your key would be sent.
Persistence & Privilege
The skill does not request 'always: true' or system-level privileges and contains no installer modifying other skills. However, SKILL.md's wording that the agent should "remember" the API key implies conversational or implicit persistence of a secret, which is a privacy risk even if not a system privilege. Prefer platform secret storage rather than having the agent retain secrets in chat history.
What to consider before installing
This skill appears to implement what it claims (querying FEEDAX for bond news) but has two practical concerns you should address before installing or using it: 1) Do not paste your API key into the chat. SKILL.md asks you to "tell me API Key so I can remember" — that would expose the secret in conversation history. Instead, supply the FEEDAX_API_KEY via a local environment variable, a local config file kept off-line, or the platform's secret storage mechanism if available. 2) The registry metadata does not declare any required credentials, yet the code requires FEEDAX_API_KEY. Ask the publisher (or the registry) to correct the manifest to list FEEDAX_API_KEY as a required credential so the platform can treat it appropriately. 3) Verify the endpoint: the script sends requests to IP 221.6.15.90:18011 rather than a named feedax domain. Confirm with FEEDAX or the publisher that this is the correct and trusted endpoint before providing keys. 4) Check dependencies locally (the script uses the 'requests' library) and run the script in an isolated environment first. If you must use the skill, prefer passing the API key via env var/CLI/config file locally and avoid sharing secrets in chat. If you'd like, I can suggest edits to SKILL.md and the manifest to remove the prompt that asks users to paste the API key and to declare the required credential explicitly.

Like a lobster shell, security has layers — review code before you run it.

latestvk97ff2j0w90gd1hdxerdq9xtns845702

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments