Skillv1.2.0

ClawScan security

SQL to BI Builder · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 12, 2026, 6:40 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's code, scripts, and runtime instructions are consistent with its stated purpose (converting SQL markdown into dashboard specs, UI scaffolds, and a demo backend) and do not request unrelated credentials or perform network exfiltration.
Guidance: This skill appears to do what it claims: parse SQL markdown and emit a dashboard spec, a static UI scaffold (which loads fonts/echarts from public CDNs), and a local FastAPI demo backend that serves synthetic preview data. Before running: ensure you use a dedicated working directory or specify an output path you control (the scripts write files and may start local services), install Python 3.11 in a virtualenv as instructed, and inspect the generated services/start scripts if you plan to run them (the demo backend enables CORS and listens on localhost by default). There are no hidden network exfiltration signals or credential requests, but avoid feeding real production credentials or sensitive SQL containing secrets into any demo outputs without reviewing how you store/share the generated files.

Review Dimensions

Purpose & Capability: okThe name/description match the actual files and behavior: parsing sql.md, inferring semantics, recommending charts, building dashboard.json, scaffolding a frontend, and emitting a demo FastAPI backend. No environment variables, external credentials, or unrelated binaries are required.
Instruction Scope: okSKILL.md directs only local actions (create venv, run the included Python scripts, write outputs to user-specified paths, and optionally start local demo services). The instructions reference repository-local files and tooling (pyproject, .python-version) for correctness; they do not instruct reading unrelated system files or sending data to external endpoints. The frontend scaffold does reference public CDNs for fonts and echarts (normal for a UI demo).
Install Mechanism: okThere is no install spec; the skill is instruction-plus-scripts and runs in a Python venv. All code is bundled with the skill; no remote downloads or archive extraction are performed by the skill itself (aside from typical CDN usage in generated frontend HTML).
Credentials: okThe skill declares no required environment variables or credentials and the scripts do not attempt to read secrets. Runtime notes ask for Python 3.11 and optional dev deps (PyYAML) which are proportional to the parsing/generation tasks.
Persistence & Privilege: okThe skill does not request always:true and will not auto-inject itself. It writes generated artifacts to an output directory specified by the user and can start local services; this is expected for a scaffold generator and does not modify other skills or system-wide agent settings.