Skillv9.9.5

ClawScan security

Schema Markup Generator · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

BenignApr 28, 2026, 4:05 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's requests and runtime instructions are consistent with its stated purpose (generating JSON-LD schema); it asks for no credentials, installs nothing, and contains no surprising I/O or exfiltration behaviors.
Guidance: This skill appears coherent and low-risk: it generates JSON-LD templates, validates them, and gives implementation guidance without requesting secrets or installing code. Before installing, consider: (1) review any outputs before publishing (ensure you don't accidentally publish PII or fabricated review/rating data), (2) confirm you are comfortable with the agent storing schema artifacts and handoff summaries in its memory/store, and (3) when using the WebFetch tool, treat fetched pages as untrusted source content — validate and sanitize any data pulled from external pages before embedding it in published markup.

Purpose & Capability: okName and description (JSON-LD/schema generation for FAQ, HowTo, Product, LocalBusiness, Article) match the SKILL.md and reference files. No unrelated env vars, binaries, or installs are required.
Instruction Scope: okThe instructions focus on identifying content type, producing JSON-LD, implementation guidance, and validation. They explicitly treat fetched web content as untrusted data and do not instruct reading system files, secrets, or unrelated configuration.
Install Mechanism: okInstruction-only skill with no install spec and no code files — lowest-risk install profile. All runtime behavior is described in prose.
Credentials: okRequires no environment variables, credentials, or config paths. Referenced web resources are external documentation and optional WebFetch usage is appropriate for the task.
Persistence & Privilege: noteSkill declares it will write schema deliverables and handoff summaries to agent memory paths (memory/content/, memory/decisions.md, etc.). This is coherent for a content-generation skill, but users should be aware outputs may be persisted to the agent's memory store.