StruktoLab Studio

AI Security Review

Evidence context: threat category none; evidence quality moderate.

What the extension is

StruktoLab Studio is a VS Code extension from OpenPatch, version 0.5.1, that draws Nassi-Shneiderman structograms for .struktolab files. It declares no permissions and no host permissions. The only code that reaches the network is a fetch call at extension/dist/webview.js:291, recorded as NET-FETCH-extension/dist/webview.js-291. That path places the call inside the extension's bundled webview output, which is where a diagram editor keeps the code that renders and refreshes its editor panel.

Is the access justified?

The three endpoints in play are struktolab.new, struktolab.openpatch.org and schemas.openxmlformats.org. Two are the publisher's own project domains. The third is the namespace host for the Open XML document format, a reference any tool that emits or validates OOXML documents has an ordinary reason to carry. Nothing in the manifest-analysis or secret categories fired, so there is no evidence of the extension reaching outside the workspace or touching process execution. This is a file editor, so reading and writing .struktolab files in the open workspace is the job, not a side effect.

Credential findings

The secret category is empty. No finding references .env, .ssh, .git/config, cloud credential stores, or VS Code's secret storage. The extension declares no permissions at all, which means there is no scope it asked for that its purpose doesn't cover.

Where the noise comes from

The 18 code-smell entries are low severity and sit on bundled JavaScript under extension/dist/. Rules tagged postinstall_* and credential_* match basic Node.js vocabulary such as fetch, exec, fs and process.env, and they fire on nearly any non-trivial bundle. Six indicator hits appear with no specific hostile domain attached; the only named endpoints are the two publisher domains and the schema namespace. There are zero malware-family, malware-signature, obfuscation and tool-poisoning matches.

Strongest counterargument

15 installs is a small audience, and this evidence does not independently confirm OpenPatch's marketplace verification. Low download count plus a recent publish is standard supply chain heuristics. But that heuristic needs either broad permissions or a hostile endpoint to mean anything, and this extension has neither: no permissions, no host permissions, and no endpoint outside the publisher's control or a standards schema host. A typosquat would also need a name close to a popular extension, and StruktoLab Studio isn't imitating anything. Neither signal survives contact with the actual findings.

Conclusion

Nothing in the code supports data theft, credential access or payload execution. The volume comes from bundled webview code and broad pattern rules, not from anything the extension does.

Key Reasons

• No permissions or host permissions declared, and no filesystem, secret or process-execution findings of any kind.
• The single network finding, NET-FETCH-extension/dist/webview.js-291, sits in bundled webview build output, which is expected for a webview-based visual editor.
• Named endpoints are the publisher's own domains (struktolab.openpatch.org, struktolab.new) and schemas.openxmlformats.org, the Open XML schema namespace.
• Zero malware, malware-signature, obfuscation, secret and tool-poisoning matches across the scan.
• The 18 code-smell entries are low-severity YARA pattern matches on bundled JavaScript under extension/dist/, not behavior.

False Positive Considerations

• 18 low-severity code-smell rules matching generic Node.js vocabulary (fetch, exec, fs, process.env) in bundled extension/dist/ JavaScript
• Six unnamed IoC hits with no specific hostile domain; every named endpoint is either publisher-controlled or the OOXML schema host
• Single NET-FETCH finding located in minified webview build output rather than source
• Finding volume concentrated in dist/ bundle files, which multiplies pattern matches per bundled library

Reviewed 2026-10-12; recommended action: suppress false positive; model confidence 85%.