Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
MigraQ
v1.10.0腾讯云迁移平台(CMG/MSP)全流程能力。触发词:资源扫描、扫描阿里云/AWS/华为云/GCP资源、生成云资源清单、选型推荐、对标腾讯云、推荐规格、帮我推荐、给我推荐、ECS对应什么腾讯云产品、成本分析、TCO、迁移报价、询价、价格计算器、cmg-scan、cmg-recommend、cmg-tco
⭐ 0· 181·0 current·0 all-time
byHaoBin@haobinaa
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name, description, required binaries (python3), required env vars (TENCENTCLOUD_SECRET_ID / TENCENTCLOUD_SECRET_KEY), and network permissions (cmg.ai.tencentcloudapi.com, msp.cloud.tencent.com) are coherent for a Tencent Cloud migration helper that calls a CMG ChatCompletions API.
Instruction Scope
SKILL.md instructs the agent/user to run local Python scripts (check_env.py and migrateq_sse_api.py) and to always run environment checks before API calls. It also prescribes writing AK/SK into shell config files using echo >> (persisting credentials in plain text). These are broader than strictly necessary for a single API call and could lead to long-lived secret exposure. The skill also enforces an identity intro on first API call, which is unusual but not harmful. Overall the runtime instructions grant the skill broad discretion about how and where credentials are stored.
Install Mechanism
No install spec (instruction-only plus bundled scripts). No remote downloads or archive extraction. Risk is limited to running the included Python scripts locally; nothing writes code to disk beyond the existing skill files.
Credentials
The skill only requests Tencent AK/SK which is appropriate for calling CMG. However, the SKILL.md explicitly recommends persisting AK/SK into ~/.zshrc (or user environment on Windows) via echo >> which writes secrets to disk in cleartext. This contradicts the front-matter security claim that AK/SK are read from env and "not written to file or log." Persisting long-lived AK/SK in dotfiles increases exposure risk; the skill does not request any unrelated credentials.
Persistence & Privilege
always is false and the skill does not request system-wide privileges or modify other skills. It requires running its own scripts but does not claim or request persistent platform-wide presence.
What to consider before installing
What to consider before installing:
- Purpose fit: This skill appears to be what it says — a Tencent Cloud migration assistant that calls the CMG ChatCompletions endpoint and needs your Tencent AK/SK.
- Main risk: SKILL.md tells users to persist SecretId/SecretKey in shell profile files (e.g., appending to ~/.zshrc). Storing long-lived AK/SK in plain text dotfiles is risky — prefer ephemeral or least-privilege credentials (short-lived STS tokens or a dedicated low-privilege account) and avoid committing dotfiles to VCS.
- Inconsistencies to verify: the skill's front matter says AK/SK are "not written to file or log", yet instructions show echo >> to ~/.zshrc. The API docs file also contains an inaccurate Authorization example (says "Bearer <TENCENTCLOUD_SECRET_KEY>" while the scripts implement TC3-HMAC-SHA256). These mismatches suggest sloppy documentation; inspect the scripts before trusting them.
- Practical recommendations:
- Review the bundled Python scripts locally (they are included) before running them. They appear to only implement TC3 signing and SSE reading, but confirm there are no hidden endpoints or exfiltration code in the truncated sections.
- Use a dedicated least-privilege TencentCloud subaccount (as the SKILL.md suggests) and prefer short-lived credentials if possible.
- Do not blindly run the suggested echo >> commands; instead set env vars in the session or use a secure secrets manager. If you must persist credentials, use OS-provided secure storage (e.g., Windows Credential Manager, macOS Keychain) or protect dotfiles and your repo.
- Run the scripts in an isolated environment (VM or container) until you validate behavior.
- If you want higher assurance: ask the skill author for an explanation of the conflicting statements about key handling and for a signed/official release URL. If the author can't clarify, treat the conflicting guidance as a red flag and avoid using persistent credentials with this skill.Like a lobster shell, security has layers — review code before you run it.
latestvk971gtg4ekszzpvw80ndxe7dd984g6q3
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🚀 Clawdis
Binspython3
EnvTENCENTCLOUD_SECRET_ID, TENCENTCLOUD_SECRET_KEY
