Back to skill
Skillv1.0.0

ClawScan security

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

Scanner verdict

BenignMar 15, 2026, 3:14 PM
Verdict
Benign
Confidence
high
Model
gpt-5-mini
Summary
This is an instruction-only analytics/tracking guidance skill that requests no credentials, installs, or binaries and is coherent with its stated purpose.
Guidance
This skill is a content/implementation guide for analytics (GA4, GTM, UTM) and appears internally consistent. Before installing: (1) confirm you are comfortable the agent may read project-local context files (the skill explicitly checks .agents/product-marketing-context.md or .claude/product-marketing-context.md) — remove or redact sensitive info there if needed; (2) do not paste or store production secrets (measurement IDs, user identifiers, or PII) into shared project files the agent can access; (3) remember code snippets are examples — if you copy them into your site, ensure they do not send PII and that consent/consent-mode rules are applied; (4) if you plan to let the agent act autonomously, consider reviewing the commands or changes it will propose before applying them to production GTM/GA accounts.

Review Dimensions

Purpose & Capability
okThe name/description (analytics, GA4, GTM, tracking plans) matches the content of SKILL.md and the included references. There are no unrelated environment variables, binaries, or install steps requested.
Instruction Scope
noteThe runtime instructions ask the agent to read local context files if present (.agents/product-marketing-context.md or .claude/product-marketing-context.md) before asking questions — this is reasonable for context but means the skill will access repository/workspace files if available. Otherwise, instructions are scoped to analytics implementation, debugging, naming, and privacy best-practices and do not direct the agent to read arbitrary system credentials or send data to unexpected endpoints.
Install Mechanism
okNo install spec and no code files to execute — the skill is instruction-only, which minimizes on-disk risk.
Credentials
okThe skill requires no environment variables, secrets, or external credentials. Code snippets reference typical GA4/GTM identifiers (e.g., measurement IDs) which are expected for the purpose but are not requested as required environment values in the registry.
Persistence & Privilege
okalways is false (default) and the skill is user-invocable; the skill does not request elevated or permanent platform privileges. Note: autonomous invocation is allowed by default on the platform but is not itself unusual for skills and is not combined here with other concerning flags.