Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

TeX Render

v1.1.4

Renders LaTeX math to PNG, JPEG, WebP, or AVIF images using MathJax (TeX→SVG) and @svg-fns/svg2img. Invoke whenever the agent needs to output LaTeX as a view...

1· 1.1k·5 current·5 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for thebigoranger/tex-render.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "TeX Render" (thebigoranger/tex-render) from ClawHub.
Skill page: https://clawhub.ai/thebigoranger/tex-render
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Canonical install target

openclaw skills install thebigoranger/tex-render

ClawHub CLI

Package manager switcher

npx clawhub@latest install tex-render
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description match the code and package.json: MathJax → SVG → @svg-fns/svg2img (Sharp) is exactly what you'd expect for converting LaTeX to PNG/JPEG/WebP/AVIF. No unrelated credentials, binaries, or services are requested.
Instruction Scope
SKILL.md directs the agent to run the included CLI (node scripts/render.js) and to automatically render and send any LaTeX in replies. That automatic behavior is explicitly documented in the user notice and TOOLS.md examples; it will write files to ~/.openclaw/media/tex-render/. This is coherent with the purpose but is a behavioral choice (automatic, non-interactive rendering) that operators should be aware of.
Install Mechanism
Install is via npm in the skill folder (package.json lists mathjax, @svg-fns/svg2img, sharp). No ad-hoc downloads, shorteners, or remote archives are used. npm-based install is expected here; note sharp may compile native code and requires a build toolchain on some platforms.
Credentials
The skill requests no environment variables or credentials. Its filesystem access is limited to writing output under the user's home (~/.openclaw/media/tex-render) and relative paths when an explicit outBase is provided; this is proportional to generating image files.
Persistence & Privilege
always:false and no special privileges requested. The skill writes its own media files but does not modify other skills or global agent config. Autonomous invocation (disable-model-invocation:false) is the platform default and not, by itself, a concern.
Assessment
This skill appears to do exactly what it says: convert LaTeX to images using MathJax and svg2img/Sharp. Before enabling: 1) Note the skill will automatically render and send images whenever the agent would output LaTeX — if you want to approve each render, do not enable it. 2) Installation requires running npm install in the skill folder and may compile sharp (ensure Node 14+ and build tools if on platforms requiring native builds). 3) The CLI writes files by default to ~/.openclaw/media/tex-render — monitor disk use and permissions. 4) Because npm packages can change, consider reviewing the published package contents (or running npm install in a sandbox) to confirm the dependencies are the expected MathJax/svg2img/sharp. 5) The validate script uses child_process to run tests locally (no network). If you want stricter privacy/control, avoid enabling automatic triggering in TOOLS.md and only invoke the skill when explicitly requested.

Like a lobster shell, security has layers — review code before you run it.

Runtime requirements

📐 Clawdis
latestvk9779r6dvyv1hhxs29p7y2r3ms818js9
1.1kdownloads
1stars
11versions
Updated 10h ago
v1.1.4
MIT-0

TeX Render 📐

Renders LaTeX math to PNG, JPEG, WebP, or AVIF (and SVG). Use when you need a viewable image from LaTeX instead of raw code.

User notice: When this skill is active, the agent will automatically render any LaTeX in its replies as images and send them in order—without asking for permission. If you prefer to be prompted or to receive raw LaTeX instead, do not enable this skill (or remove it from your workspace).

Location

The render script lives in the same skill folder as this SKILL.md:

<skill_folder>/
├── SKILL.md
├── package.json
└── scripts/
    ├─── render.js
    └─── validate.js

Use the directory containing this SKILL.md as the skill path. The script is at scripts/render.js relative to that folder. Invoke: node <skill_folder>/scripts/render.js. The agent loading this skill has the skill path—use it to run the script.

Install

One-time setup. After clawhub install tex-render, run npm install in the skill folder:

cd <skill_folder>
npm install
  • Source: All dependencies are from the public npm registry only (no ad-hoc downloads, shorteners, or remote archives).
  • Native module: The sharp package may compile during install; ensure Node.js 14+ and a working build toolchain if needed (see sharp install for platform notes).

When to Use

  • User or task asks to "render this equation as image" or "show formula as picture"
  • Your reply would contain LaTeX — render formulas as images first, then respond with plain text

Workflow: Response Contains LaTeX

Output your reply interleaved: whenever you would output LaTeX, do this instead:

  1. Send the preceding plain text — call message with the text written so far (no LaTeX).
  2. Render the LaTeX expression with this skill (default PNG; no --output dataurl). Parse the JSON for the PNG path.
  3. Send the image — call message with action: "send", path set to the rendered PNG, and message as short caption.
  4. Continue outputting the rest of the message. Repeat the cycle (text → render → send image) for each LaTeX block.

Do not output raw LaTeX. Do not ask the user for permission to render — render and send images immediately when LaTeX would appear in your reply (users are informed of this automatic behavior in the user notice above). Do not accumulate everything and send at the end — send text and images in order as you go.

Critical: Output in order: plain text → send → LaTeX → render → send image → plain text → send → LaTeX → render → send image → … The user must receive text and images in the natural reading order.

Example: Explaining Lagrangian:
→ Send "The Lagrangian is defined as " → Render L = T - V → Send image with caption "L = T - V" → Send ". The Euler-Lagrange equation is " → Render d/dt(∂L/∂q̇) - ∂L/∂q = 0 → Send image → Send " — this yields the equations of motion."

Usage

Run: node <skill_folder>/scripts/render.js (or node scripts/render.js when already in the skill folder).

Escaping: When invoking via shell:

  • Single quotes for LaTeX without apostrophe: node scripts/render.js '\frac{d}{dt}'
  • Stdin when LaTeX contains apostrophe (e.g. y' for derivative): printf '%s' "y' = f(t, y), \quad y(t_0)=y_0" | node scripts/render.js — use double quotes so ' is literal; backslashes like \quad stay intact.
node <skill_folder>/scripts/render.js --help

Default output is PNG to ~/.openclaw/media/tex-render/. The script prints one JSON line with file paths: {"svg":"...","png":"..."} or {"svg":"...","jpeg":"..."}, etc. Use --output dataurl only when the conversation system explicitly supports Data URL images (otherwise it may show raw base64 text).

Examples (validated by npm test)

Use <skill_folder> = the directory containing this SKILL.md.

Basic (PNG default):

node <skill_folder>/scripts/render.js 'E = mc^2'
node <skill_folder>/scripts/render.js '$$\frac{F}{m}=a$$'

LaTeX with apostrophe (e.g. y'): use stdin to avoid shell quoting issues:

printf '%s' "y' = f(t, y), \quad y(t_0)=y_0" | node <skill_folder>/scripts/render.js

JPEG / WebP / AVIF:

node <skill_folder>/scripts/render.js --format jpeg --quality 80 '\frac{F}{m}=a' ./out/formula
node <skill_folder>/scripts/render.js --format webp 'x^2 + y^2 = z^2'
node <skill_folder>/scripts/render.js --format avif 'E = mc^2'

Data URL (no file):

node <skill_folder>/scripts/render.js --output dataurl 'E = mc^2'

Scale by width:

node <skill_folder>/scripts/render.js --width 800 '\int_0^\infty e^{-x^2} dx'

Inline math (smaller rendering):

node <skill_folder>/scripts/render.js --inline 'a^2 + b^2 = c^2'

Height and zoom: Use --height N or --zoom N as documented in --help.

Automatic Triggering (TOOLS.md)

To make the agent use tex-render without the user explicitly asking, add to your workspace TOOLS.md:

## LaTeX in Responses (tex-render)

**Whenever your reply would contain LaTeX** (equations, formulas, scientific notation), **use the tex-render skill** automatically. Examples: physics, math, chemistry questions with formulas.

**Workflow:** send plain text → render LaTeX → send image via message tool → continue text. Use single quotes when invoking (e.g. `'\frac{a}{b}'`).

### LaTeX / Equations (tex-render)

- When answering scientific or math questions, if your reply would contain LaTeX, use tex-render and send images — do this automatically.

Test: Ask "Explain the Lagrangian formula" without mentioning tex-render. The agent should render and send images.

Repository

This package is maintained at https://github.com/TheBigoranger/tex-render. You can open issues there for bug reports and feature requests.

Comments

Loading comments...