Codebase to Course
AdvisoryAudited by VirusTotal on Apr 15, 2026.
Overview
Type: OpenClaw Skill Name: codebase-to-course-generator Version: 1.0.0 The codebase-to-course skill bundle is a legitimate tool designed to generate interactive educational HTML content from source code. The SKILL.md instructions guide the agent through codebase analysis, curriculum design, and file assembly using a simple shell script (build.sh) and a comprehensive JavaScript engine (main.js) for UI interactivity. While it performs potentially sensitive actions like cloning remote repositories (git clone) and reading local files, these are strictly aligned with its stated purpose, and no evidence of data exfiltration, malicious persistence, or prompt-injection attacks was found.
Findings (0)
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.
The agent may download a repository and inspect many files in the selected project.
The skill asks the agent to use git and read a local or cloned repository. That is expected for codebase analysis, but it is still a broad user-directed file/source operation.
If the user provides a GitHub link, clone the repo first (`git clone <url> /tmp/<repo-name>`) before starting the analysis. If they say "this codebase" or similar, use the current working directory.
Use it only on repositories you intend to analyze, and avoid invoking it from a directory that contains unrelated private files.
If the source project contains secrets, internal logic, or proprietary code, those details could end up in the generated HTML course.
The generated course can persist exact source-code excerpts. This is central to the skill’s educational purpose, but users should treat the output as potentially containing private or proprietary code.
Code snippets are exact copies from the real codebase — never modified or simplified. The learner should be able to open the actual file and see the same code they learned from.
Review the generated course before sharing it, and avoid running the skill on projects containing secrets unless you first remove or redact them.
A poorly generated course could render unintended HTML inside quiz feedback.
The course runtime inserts generated quiz explanation text as HTML. This appears intended for locally generated course content, not remote input, but generated explanations should be escaped or trusted to avoid accidental HTML/script injection.
feedback.innerHTML = '<strong>Exactly!</strong> ' + rightExp;
Prefer escaping generated text or using textContent for untrusted content, especially if course text is derived from untrusted repositories.
Opening the generated course may contact Google font servers.
The generated course loads Google Fonts from an external CDN. SKILL.md discloses this as the only external dependency, but it means the opened course is not fully self-contained unless fonts are removed or vendored.
<link rel="preconnect" href="https://fonts.googleapis.com">
If offline use or privacy is important, self-host the fonts or remove the external font links before distributing the course.