Webnovel Serial Pipeline

Build and publish a Quartz-hosted Korean web-novel serial (draft→cover→webp→episode md→lint→publish). Uses Nano Banana Pro + ffmpeg.

This package is mostly coherent with its stated purpose (local authoring → lint → copy → index update → image conversion). Before installing, check these things: (1) update/confirm metadata: python3 and ffmpeg are required and WEBNOVEL_QUARTZ_ROOT is mandatory — the registry should list them; (2) inspect scripts yourself (they are small and readable) to confirm they will only operate on the directories you point at; (3) if you plan to use cover-generation, confirm how NANO_BANANA_KEY is used and only provide that key if you trust the Nano Banana Pro service; (4) run python3 scripts/prepublish_check.py in the skill folder to verify nothing unexpected is present; (5) run publish_workflow on a test copy of your Quartz vault to confirm behavior (the default is to copy, not delete drafts). The main risks here are metadata/documentation mismatches (surprising requirements) and possible AV false positives from the prepublish_check signatures — not evidence of remote exfiltration. If you want higher assurance, ask the publisher to correct the registry fields or provide a signed release.

Type: OpenClaw Skill Name: webnovel-serial-pipeline Version: 0.1.4 The skill bundle is benign. All scripts (Python and Bash) perform local file operations, content linting, dependency checks, and image conversions, which are directly aligned with its stated purpose of building and publishing web novels. Notably, the `SKILL.md` explicitly addresses potential antivirus false positives and clarifies that `prepublish_check.py` contains regex patterns (e.g., for `curl`, `wget`, `webhook.site`, private keys) to *detect* malicious code, not to execute it. This `prepublish_check.py` script acts as a security feature, actively scanning for common exfiltration and remote execution markers within the skill's own files, demonstrating a proactive security posture rather than malicious intent.