Skillv1.0.0

ClawScan security

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

BenignMar 11, 2026, 11:37 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's files, instructions, and requirements are coherent with its stated purpose (generating and validating ShrimpCard JSON and rendering HTML) and do not request unrelated credentials, network access, or unusual installs.
Guidance: This skill appears to do exactly what it says: collect required fields, validate a ShrimpCard JSON, and render a standalone HTML preview with a pixel-art fallback. Before installing, be aware it will read/write files in the agent's working directory (shrimp-card.json and shrimp-card.html by default). It has no network code and requests no credentials, so nothing here will exfiltrate secrets. One small behavior to note: the validator enforces that lobster_image_desc contains either the Chinese character '虾' or the substring 'lobster' — this is a strict check that may cause valid image descriptions to be rejected unless they match that pattern. If you plan to use external image-generation services, the skill will not call them for you; you must provide image URLs/data URLs yourself or accept the pixel-art fallback.

Purpose & Capability: okName/description match included files: schema, sample, validator, HTML template, and render script. No unexpected env vars, binaries, or external services are requested.
Instruction Scope: noteSKILL.md instructs the agent to gather fields, validate JSON via scripts/validate_card.py, and render HTML via scripts/render_card_html.py. These steps operate on local files and write shrimp-card.json and shrimp-card.html to disk as intended. Note: the skill will read and write files in the agent's working directory (normal for this purpose).
Install Mechanism: okInstruction-only with included Python scripts and assets; no install spec, no external downloads, and no package installation. Low-risk from install standpoint.
Credentials: okNo environment variables, credentials, or config paths are required. The skill does not request unrelated secrets or system access.
Persistence & Privilege: okalways:false (no forced inclusion). The skill can be invoked by the agent autonomously (default), which is expected for a user-invocable tool; it does not modify other skills or system settings.