Skillv1.0.2

ClawScan security

News Sentiment · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

ReviewMar 6, 2026, 5:13 AM

Verdict: Review
Confidence: medium
Model: gpt-5-mini
Summary: The skill mostly matches a crypto-news sentiment tool, but there are several inconsistencies and undeclared credential requirements that make its intentions and resource needs unclear — review before installing or providing keys.
Guidance: Before installing or supplying API keys: (1) Note the skill will send tweet/news text to two external services (ai.6551.io and dashscope.aliyuncs.com); confirm you trust those domains. (2) The package omissions are concerning — TWITTER_TOKEN is required by the code but not declared in registry metadata. Only provide keys that are scoped and easily revoked. (3) Prefer testing in an isolated environment or container, and inspect/modify the Python script if you want to limit what is sent externally (for example, remove or sanitize PII). (4) If you need the 'news sites' or 'Reddit' features, ask the author for a complete, up-to-date implementation and clear documentation of required credentials and endpoints. (5) If you remain unsure, avoid providing real API tokens and run with mock/test tokens first.

Review Dimensions

Purpose & Capability: concernThe skill claims to scan Twitter, news sites, and social media, but the included code only implements a Twitter search + LLM sentiment call; README/SKILL.md mention other sources (news RSS, Reddit) that are not implemented. The registry metadata declares only DASHSCOPE_API_KEY as required, yet the code uses a TWITTER_TOKEN environment variable as well. The declared required binary 'curl' is listed but never used in the shipped Python code. These mismatches suggest sloppy packaging or missing declarations.
Instruction Scope: concernRuntime behavior (sentiment_analyzer.py) performs network calls to third-party endpoints (ai.6551.io for Twitter search and dashscope.aliyuncs.com for LLM analysis) and will send tweet/news text to the remote LLM service. SKILL.md/README mention these sources but do not declare that user data will be transmitted to these specific domains or explain privacy implications. The instructions and metadata fail to fully enumerate all required env vars (TWITTER_TOKEN is used but not listed), expanding the actual scope of data sent off-host.
Install Mechanism: okNo install spec is provided (instruction-only install), so nothing arbitrary is downloaded at install time. The only shipped code is a Python script which will run on the host; no archive downloads or remote installers are used.
Credentials: concernThe registry requires only DASHSCOPE_API_KEY, but the code also requires TWITTER_TOKEN to query the ai.6551.io Twitter proxy — this secret is referenced in README but not declared in requires.env. The skill therefore underdeclares secrets it needs. The endpoints are third-party (including a non-official Twitter proxy domain), so providing these credentials would grant access to external services not fully described in metadata.
Persistence & Privilege: okThe skill does not request always:true, does not modify other skills or system configuration, and has no install-time persistence mechanism. Autonomous invocation is allowed (platform default) but not combined with other high privileges.