Skillv1.0.0

ClawScan security

Ai Backendfrontend Autowriter · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignMar 8, 2026, 3:36 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's requirements and instructions are consistent with its stated purpose and it requests no credentials or system access, although the included script is only a simple template runner and does not actually implement full code generation.
Guidance: This skill appears coherent and low-risk: it asks for no credentials, performs no installs, and its code is a harmless template runner. However, note that the provided run.py does not actually generate or validate code — it's a stub that prints a pipeline summary. If you plan to use this skill for real code generation or allow it to perform external actions (deployments, network calls), review and extend the implementation yourself and test in a sandbox. Also confirm the agent enforces the SKILL.md's secondary-confirmation requirement before any external or destructive operation. (Minor note: the run.py imports UTC from datetime which may not exist in some Python versions — that's a runtime bug, not a security issue.)

Purpose & Capability: noteName/description: generate and validate backend/frontend code. Declared requirements: none. Actual artifacts: SKILL.md describes a reasonable pipeline for that purpose, but the shipped run.py merely prints a template JSON summary and does not perform real code generation or validation — this is a capability gap (not a security problem) but is inconsistent if the user expects a fully implemented generator.
Instruction Scope: okSKILL.md workflow is scoped to analyze, plan, execute by role, validate, and return a summary. It explicitly requires secondary confirmation before external actions/deployments and says not to perform destructive actions unless authorized. The instructions do not request reading unrelated files, environment variables, or external endpoints.
Install Mechanism: okNo install spec and no downloads. Instruction-only with a small included script — nothing is written to disk by an installer and there are no third-party packages fetched.
Credentials: okThe skill requires no environment variables, credentials, or config paths. The run.py does not access environment variables or network resources. There is no disproportionate credential request.
Persistence & Privilege: okSkill is not always-enabled and does not request persistent or system-wide privileges. There is no behavior that modifies other skills or system configuration.