Math Worksheets

PassAudited by ClawScan on Feb 24, 2026.

Overview

The skill's code, files, and runtime instructions are coherent with its stated purpose (generating LaTeX worksheets, verifying with SymPy, and compiling with tectonic); no unrelated credentials or network installs are requested, though a couple of expected-but-notable behaviors merit attention.

This skill is coherent with its stated purpose, but review these practical points before installing: - SymPy evaluation: verify.py uses sympy.sympify/nsimplify to parse expressions from a generated JSON. This is necessary to check math, but sympify is a parsing/evaluation surface — avoid running verification on JSON from untrusted third parties without review. If you accept uploaded verification files, audit the inputs first or run verification in a sandbox. - Tectonic downloads: compiling LaTeX with tectonic may cause the tool to fetch TeX packages from the network on first use. This is expected but means transient network activity when compiling PDFs. - Config reading: check_reasoning_model.sh optionally reads ~/.openclaw/config.json to detect configured models. It extracts model IDs (not tokens) to decide which model to use; if your OpenClaw config contains secrets in unusual fields, consider moving them or running the skill in an environment where that config is benign. - Delivery claims vs implementation: README mentions "channel mirroring" delivery (Telegram, iMessage, email) but the provided scripts do not implement network-based delivery; ensure your deployment/integration handles output delivery securely rather than relying on undocumented behavior. Recommendations: inspect the included scripts locally, ensure python3 + sympy and tectonic are installed from trusted sources, and run the first few worksheet generations in a sandboxed user account to verify behavior and network activity. If you need higher assurance, a quick code review focused on sympify usage and any future changes to network calls is advisable.