Back to skill
Skillv1.0.0
ClawScan security
Ads Creative Ideas · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 3, 2026, 3:49 AM
- Verdict
- Benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- This instruction-only skill is internally consistent with its stated purpose (generating ad creative ideas) and does not request unexpected credentials, installs, or system access.
- Guidance
- This skill appears coherent and low-risk, but remember: (1) do not share sensitive business credentials, customer PII, or proprietary creative assets unless you trust the environment; (2) treat suggested strategies and metrics as recommendations — validate with platform-specific account data and policy checks before launching; (3) test any recommendations on small budgets or in staging/ad previews first; (4) if you want to limit autonomous invocation, disable model-driven invocation in agent settings or require manual invocation.
Review Dimensions
- Purpose & Capability
- okName/description (ads creative inspiration across platforms) aligns with the SKILL.md content. Required inputs, outputs, workflow, and platform notes are all relevant to creative planning for Meta, TikTok, YouTube, Google, and Amazon ads. There are no surprising environment or binary requirements that don't belong to this domain.
- Instruction Scope
- okSKILL.md contains purely guidance for generating creative briefs, angle/hook libraries, variant matrices, QA checks, and escalation rules. It does not instruct the agent to read local files, access system paths, transmit data to third-party endpoints, or use credentials outside the declared (none). Guardrails explicitly discourage fabricating metrics and require clear separation of facts vs assumptions.
- Install Mechanism
- okNo install spec and no code files — the skill is instruction-only, so nothing is written to disk or downloaded. This is the lowest-risk installation model and is proportional to the skill's purpose.
- Credentials
- okThe skill requires no environment variables, credentials, or config paths. That is appropriate for an instruction-only creative generation skill that provides planning and copy/asset guidance rather than interacting directly with ad platforms or accounts.
- Persistence & Privilege
- noteDefaults allow agent invocation and autonomous use (disable-model-invocation: false) but always:false. Autonomous invocation is the platform default and not inherently problematic here; users should be aware the agent could call the skill when relevant prompts arise.