Back to skill
Skillv1.0.0
ClawScan security
seo-audit · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
SuspiciousMar 11, 2026, 3:51 PM
- Verdict
- suspicious
- Confidence
- medium
- Model
- gpt-5-mini
- Summary
- The skill is generally coherent for performing on‑page SEO checks, but the runtime instructions tell the agent to read a local product-context file that is not declared and the included analyzer is limited to on‑page checks (other tasks require external tools or user-supplied credentials), so there are some scope/visibility mismatches you should review before installing.
- Guidance
- This skill appears to be a legitimate SEO-audit helper, with a bundled Python on‑page analyzer. Before installing or invoking it: (1) check for and inspect any local .claude/product-marketing-context.md file—do not store secrets there because SKILL.md tells the agent to read it but the skill metadata did not declare that path; (2) be cautious about providing Search Console or analytics credentials—only supply them if you trust the environment and understand what will be accessed; (3) the bundled script may fetch a URL you give it (urllib.request) to analyze pages—avoid giving private/staging URLs unless you want the agent host to access them; (4) if you want only offline analysis, don’t provide URLs or ask the agent to run the script against local files only; and (5) if concerned, review scripts/seo_checker.py yourself (it’s included) to confirm behavior before running.
Review Dimensions
- Purpose & Capability
- noteName/description align with the provided assets: SKILL.md describes a full SEO audit and includes a Python on‑page analyzer (scripts/seo_checker.py). The script implements on‑page checks (title, meta, headings, images, links, word count, viewport) which matches part of the declared audit scope. However, the SKILL.md also promises broader technical/crawlability/speed checks (Core Web Vitals, sitemap, Search Console analysis) that are not implemented in the script and are delegated to external tools; that difference is explainable but worth noting.
- Instruction Scope
- concernSKILL.md instructs the agent to read .claude/product-marketing-context.md if it exists before asking questions, but the skill metadata did not declare that config path. The instructions also prompt the agent to ask for Search Console/analytics access when appropriate—this is reasonable for an audit but it means the agent may request or expect credentials or private URLs. The instruction to read an undeclared local file is a mismatch and could cause the agent to access user data without that being obvious in the skill requirements.
- Install Mechanism
- okNo install spec (instruction-only) and included code file only; no external downloads or install commands. The Python script is bundled with the skill; there is no installer that pulls remote code, which lowers install-time risk.
- Credentials
- noteThe skill declares no required environment variables, credentials, or config paths. That is consistent with the bundled on‑page analyzer. However, SKILL.md references use of Search Console/analytics and may prompt the user for access or tokens at runtime—requests for such credentials are expected for a full audit, but you should not provide credentials unless necessary and understood.
- Persistence & Privilege
- okThe skill does not request always:true and uses default invocation settings. It does not claim to modify other skills or persist system-wide settings. Normal autonomous invocation is allowed (default), which is expected for skills.