making-of-film

Warn

Audited by Socket on Nov 21, 2026

4 alerts found:

Anomalyx4
AnomalyLOW
scripts/explainer.js

This browser animation library renders scenes and captions, with no apparent malware, credential theft, or suspicious network destination. Caption values from o.words reach innerHTML without escaping, creating a potentially serious XSS risk when those values are attacker-controlled; root.CAM_HTML is also interpreted as HTML. Playback time is sent to the parent window with a wildcard origin, and o.voice may initiate expected media playback. Overall severity is moderate and depends on whether configuration values are trusted.

Confidence: 99%Severity: 62%
AnomalyLOW
scripts/film-facts.mjs

This utility inspects a local film HTML page and prints its timing and chapter metadata. It executes the page's scripts in headless Chromium with `--allow-file-access-from-files`, so using it on untrusted HTML can expose local files or permit page-initiated network activity; no direct exfiltration destination, credential targeting, or malicious payload appears in this code. Severity is conditional on the input HTML being untrusted; no external domains, URLs, or IP addresses are hardcoded.

Confidence: 97%Severity: 62%
AnomalyLOW
scripts/preview.sh

This is a media-preview encoding script with no evident network communication, credential collection, or destructive behavior beyond overwriting its requested output. It has exploitable input-handling risks if HTML content or environment variables are untrusted: crafted scene IDs can alter the FFmpeg concat list, and BED_AT can inject Python source. Severity is moderate in that trust scenario; no malicious intent is evident from the supplied code.

Confidence: 97%Severity: 58%
AnomalyLOW
scripts/fix-onsets.mjs

This is a data-maintenance script that adjusts word timings and writes the results to caller-specified files. Its significant security risk is executing the supplied JavaScript file with new Function, which permits arbitrary code execution if that input is untrusted; the file writes can also overwrite the specified paths. No evidence of intentional malware, data exfiltration, or obfuscation is present.

Confidence: 99%Severity: 62%
Audit Metadata
Analyzed At
Nov 21, 2026, 01:56 AM
Package URL
pkg:socket/skills-sh/mengto%2Fskills%2Fmaking-of-film%2F@6ddf1dd1ad0d90ad70d3d91481a80d4874694b97