dbs-content-system turns a large folder of old writing into a structured local project you can keep adding to. It first audits the folder for volume, sources and scope, and stops with a lighter path if there is too little material. When the audit passes, it creates a new project directory, copies the originals into it, and extracts a first batch of content units from three to five sample documents.
Each unit is one Markdown file of a fixed type: question, concept, opinion, case or solution, with source and relation fields. The project also gets topic maps, topic assembly drafts, link and duplicate indexes and a status overview, with Obsidian-style links. Work moves from audit to sample, batch and full run only when each gate passes.
What You Get
- An audit with material size, what is included and excluded, and a pass or a lighter fallback
- A new project folder with numbered sections for rules, raw copies, units, status, templates, topic maps, topic assembly and scripts
- A first set of content units with type, source documents, keywords and relations
- Topic maps, at least two topic assembly drafts, link and duplicate indexes and a status overview
- A rule that originals are copied and never changed
- Next-batch guidance with checks on field, relation and duplicate rules
Common Use Cases
- Organizing years of articles, tweets and course scripts into reusable material
- Building a note graph you can open in Obsidian
- Letting an agent draft new topics from existing units with sources attached
- Deciding whether a content archive is large enough for this approach
What Makes It Unique
Most note-organizing skills tidy a few files. This one is built for large archives, refuses to start when the material or scope is too thin, and moves up through four modes only when listed gates pass, such as stable unit rules, zero missing link targets and spot checks of thirty units. It limits itself to five unit types, four relation types and four duplicate types.
Under The Hood
The agent follows SKILL.md and uses bundled Node.js scripts and templates. It checks the Node version with a shell command, then runs scripts that create folders, copy the raw files and bundled templates into a new project, and write index, status, unit and link files. The scripts read and write local files only, and the Skill makes no web requests. It needs Node.js 18 or newer and edits only the new project, never your original files.