Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Architecture Governance & Assessment
v1.4.0Architecture governance and assessment tool. Evaluate cloud architectures against best practices and generate actionable improvement reports.
⭐ 0· 52·0 current·0 all-time
byTencent CloudQ Team@2513483494
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The skill's name/description (Tencent Cloud architecture assessment) aligns with its code and runtime actions: it signs and calls advisor.tencentcloudapi.com and offers role creation for console SSO. Required env vars (TENCENTCLOUD_SECRET_ID / TENCENTCLOUD_SECRET_KEY) and python3 are expected. Minor mismatch: the repo contains publish/ClawHub guidance and other docs not required for normal runtime (extraneous but not necessarily malicious).
Instruction Scope
SKILL.md instructs users to permanently write long‑lived SecretId/SecretKey into shell rc files (e.g., ~/.bashrc) which increases credential exposure; check_env.py also runs a remote version check using the external `clawhub` command and attempts network calls. Role creation and other IAM write operations are separated into scripts that claim to require explicit consent, which is good in principle, but the guidance to persist credentials and to store role names/ARNs in ~/.tencent-cloudq increases the attack surface.
Install Mechanism
There is no remote download/install step (instruction-only with included scripts), so nothing arbitrary will be fetched/installed by the package itself. All code is bundled in the skill, reducing supply‑chain download risk.
Credentials
Requesting Tencent Cloud API keys is proportionate for this skill. However: (1) SKILL.md insists on persisting long‑lived AK/SK in shell rc (unnecessary and risky — temporary STS tokens or local in-session exports are safer); (2) the role creation description references policies (e.g., QcloudTAGFullAccess, QcloudAdvisorFullAccess) whose scope is described inconsistently across files (sometimes 'read-only', other times 'full read/write'), so you should verify the exact policies the create_role.py will attach before consenting.
Persistence & Privilege
The skill does not request always:true and does not appear to modify other skills or system-wide agent settings. It will create and use a local config directory (~/.tencent-cloudq) to save role ARN and metadata (declared). The SKILL.md guidance asking users to write credentials to shell rc would make credentials persist across sessions — that is a user-driven persistence change and increases risk; the skill itself does not forcibly enable permanent presence.
What to consider before installing
Before installing or running this skill:
- Treat long‑lived AK/SK carefully: do not blindly follow the SKILL.md instruction to append SecretId/SecretKey to ~/.bashrc/.zshrc. Prefer using ephemeral STS tokens or exporting credentials only for the session where you run the tool. If you must use long‑lived keys, rotate them often and restrict their permissions.
- Inspect create_role.py and scripts/cleanup.py to confirm the exact IAM actions and which policies will be attached. Do not consent to role creation until you understand the attached policies (ensure they follow least privilege).
- Note an undeclared dependency: check_env.py calls the external 'clawhub' command for version checks. If you don't have or trust that tool, run check_env.py with --skip-update or run scripts in an isolated environment.
- Run initial checks in a safe/test account (not production) to observe what APIs are called and what the role creation would do. Consider running in an isolated VM/container.
- If you decide to proceed, avoid persisting secrets in shell rc; instead temporarily export them in-session or use a secure secrets manager. After use, clear any environment variables and inspect ~/.tencent-cloudq/config.json for what was written (it should contain only the role ARN per documentation).
- If anything seems inconsistent (policy names, wording about read/write vs read-only), ask the publisher for clarifications or reject the skill. The skill's source/homepage is unknown — that reduces trustworthiness; prefer skills from known publishers or with public source repositories you can audit.Like a lobster shell, security has layers — review code before you run it.
latestvk97bs23bhsz4sgqab373q60fes83h8fp
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
☁️ Clawdis
Binspython3
EnvTENCENTCLOUD_SECRET_ID, TENCENTCLOUD_SECRET_KEY