Json Render Table

PassAudited by ClawScan on May 1, 2026.

Overview

This instruction-only skill appears aligned with rendering table images, but users should notice that it may install external rendering tools, run local shell/Python commands, and optionally delegate rendering to another agent/model.

This skill looks coherent for rendering PNG table screenshots. Before installing, make sure you are comfortable with npm/Playwright tool installation, local shell execution, temporary /tmp image outputs, and optional delegation to another agent/model; avoid delegation for sensitive table data.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Installing the skill may lead the agent to download and install npm/Playwright components if they are not already present.

Why it was flagged

The skill tells the agent to install external, unpinned tooling at use time. This is directly related to the stated rendering purpose, but it expands the supply-chain surface.

Skill content
If missing, run `npm i -g json-render-cli`; if Chromium is missing, run `npx playwright install chromium`.
Recommendation

Use a trusted environment, consider pinning or preinstalling reviewed versions of json-render-cli and Playwright Chromium, and avoid running unexpected package installs in sensitive environments.

What this means

The agent may run local commands to calculate layout, build JSON, and render an image file.

Why it was flagged

The provided starter template runs shell and Python logic, then invokes the renderer to write a PNG. This is expected for the rendering workflow, but it is still local command execution.

Skill content
eval "$(python3 - <<'PY' ... )" ... "$JSON_RENDER_CMD" ... -o "$OUT_PATH"
Recommendation

Run it only in an environment where local shell execution is acceptable, and keep table input to data you are comfortable processing through the renderer.

What this means

If the table contains sensitive information, that data could be exposed to a delegated agent/model during rendering.

Why it was flagged

The skill allows rendering work to be handed to another agent/model. The artifact adds handoff constraints, but table content may still be processed outside the main agent.

Skill content
Delegate rendering to a sub-agent only when output-path handoff is explicit and deterministic. ... route simple deterministic rendering to a lower-cost fast model (for example, `gemini3flash`)
Recommendation

For sensitive tables, keep rendering in the main agent and avoid delegation or alternate model routing unless the user explicitly accepts that data flow.