Back to skill
Skillv1.2.0

ClawScan security

Weibo OpenClaw Ops · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 5, 2026, 4:37 PM
Verdict
benign
Confidence
medium
Model
gpt-5-mini
Summary
The skill's instructions and requirements are coherent with a Weibo browser-automation purpose, but it stores session state on a remote runtime and recommends periodic automated keepalive checks — you should confirm storage, retention, and scheduling policies before use.
Guidance
This skill appears to do what it says: server-side browser automation for Weibo. Before installing, consider: 1) Session sensitivity — the .state/weibo-auth.json file will hold cookies/authorization tokens on the remote runtime; confirm who can access that host, how long the file is kept, and how to delete or revoke it. 2) Keepalive scheduling — the skill recommends automated checks every 6–12 hours; clarify whether those runs require explicit consent each time or will run autonomously, and ensure this won't violate Weibo's terms or trigger anti-bot protections. 3) Mutations require explicit user intent per the instructions, but enforce this in your agent configuration (don’t rely solely on prose guardrails). 4) Because the skill is instruction-only and has no provenance or homepage, prefer running it in a controlled environment (limited-access runtime, audit logging, short retention) or request the skill author to provide an origin and a privacy/retention statement. If you need higher assurance, ask for: where session files are stored, who has access, retention policy, and an option to disable automatic keepalive.

Review Dimensions

Purpose & Capability
okName/description match the instructions: the SKILL.md exclusively describes server-side browser automation (login-by-QR, persist session state, read feed/messages, publish posts, bounded like/follow workflows). Nothing in the instructions requests unrelated credentials, binaries, or installs.
Instruction Scope
noteInstructions remain focused on Weibo automation and give concrete agent-browser commands and guardrails (explicit user consent for mutations, limits for batch likes, avoid pinned posts). They do instruct creating and loading a persistent session file (.state/weibo-auth.json) and recommend periodic (6–12h) read-only keepalive checks — both are within scope but expand the runtime behavior beyond a one-off user action.
Install Mechanism
okThis is an instruction-only skill with no install spec or code files, which minimizes on-disk risk; nothing is downloaded or installed by the skill itself.
Credentials
noteThe skill declares no environment variables or external credentials, which matches its metadata. However, it explicitly instructs persisting session state (cookies/tokens) to .state/weibo-auth.json on the remote runtime — that file effectively contains authentication material and should be treated like sensitive credentials even though it's not represented as a required env/config in the metadata.
Persistence & Privilege
okThe skill does not request always:true and does not modify other skills or system settings. It does recommend periodic keepalive checks; this implies recurring network activity and persisted auth state but does not by itself indicate excessive privilege on the platform.