obsidian-bases
by kepanoobsidian-bases helps create and edit Obsidian .base files with filters, formulas, summaries, and table, cards, list, or map views. Best for turning note metadata into working knowledge base views, with YAML validation and function-reference guidance from the skill files.
This skill scores 82/100, which makes it a solid directory listing: repository evidence shows a real, reusable workflow for creating and editing Obsidian `.base` files with filters, formulas, and multiple view types, and it gives agents enough schema and validation guidance to reduce guesswork versus a generic prompt. Directory users should still expect to test results inside Obsidian because the skill is documentation-driven rather than tool-backed.
- Strong triggerability: the description explicitly maps to `.base` files, Bases, table/card views, filters, and formulas in Obsidian.
- Good operational substance: SKILL.md provides a concrete workflow, YAML schema structure, validation checks, and common error patterns.
- Useful agent leverage: the included functions reference materially helps agents build formulas without inventing syntax.
- No install command or executable helper files; adoption depends on reading and manually applying the markdown guidance.
- Validation ultimately relies on opening the `.base` file in Obsidian, so some trial-and-error remains for rendering or syntax issues.
Overview of obsidian-bases skill
What obsidian-bases does well
obsidian-bases helps an agent create and edit Obsidian .base files for database-like note views. It is most useful when you need working YAML for Bases features such as filters, formulas, summaries, and view layouts like table, cards, list, or map.
Best fit for Knowledge Bases workflows
This obsidian-bases skill is a strong fit for people managing a vault as a knowledge base, project tracker, reading list, or content dashboard. The real job-to-be-done is not “write YAML,” but “turn my notes and metadata into a usable view that actually renders in Obsidian.”
Why use this instead of a generic prompt
The differentiator is structure. The skill gives the agent a clear workflow: define scope with filters, add optional formulas, configure views, then validate YAML and property references. That reduces the common failure modes of ordinary prompting: invalid .base syntax, broken formula references, and views that do not match the vault schema.
How to Use obsidian-bases skill
Install context and what to read first
There is no special install command exposed in the skill itself; this is a guidance skill inside kepano/obsidian-skills. Start with skills/obsidian-bases/SKILL.md, then read skills/obsidian-bases/references/FUNCTIONS_REFERENCE.md. Read the workflow section first, then the schema examples, then the function reference only for formulas you actually need.
What input the obsidian-bases skill needs
For good obsidian-bases usage, give the agent:
- the goal of the base, such as “show active projects”
- sample note paths or folders
- relevant frontmatter properties, tags, and date fields
- desired view type:
table,cards,list, ormap - any computed fields you want, such as overdue status or reading time
- display expectations, such as property order or summary fields
Weak input: “Make me a base for tasks.”
Strong input: “Create a .base file for notes in Projects/, include only notes where status != done, show title, status, due, and a formula is_overdue, default to table view, and add a cards view for mobile.”
Turn a rough goal into a prompt that works
A practical obsidian-bases guide prompt should ask for a complete .base file plus validation checks. Example:
“Using obsidian-bases, create a valid .base YAML file for notes in Areas/Research/ tagged #paper. Add filters for only unread items from 2024 onward, define formulas for age_days and is_recent, create a table view ordered by title, author, year, and formula.age_days, and explain any quoting needed to avoid YAML errors.”
This works better because it specifies scope, metadata, formulas, display order, and asks for syntax-aware output.
Practical workflow and output checks
Use this sequence:
- Ask for the minimal
.basefirst. - Validate YAML syntax before adding complexity.
- Add one formula at a time.
- Add a second view only after the first renders.
Important quality checks from the skill:
- quote strings when YAML special characters may break parsing
- ensure every
formula.Xused in a view is defined informulas - confirm property names match your notes exactly
- test in Obsidian, not just in a text editor
obsidian-bases skill FAQ
Is obsidian-bases good for beginners?
Yes, if you already understand basic Obsidian properties and frontmatter. The obsidian-bases skill is easier than learning the whole format from scratch because it gives a safe creation order. Absolute beginners may still struggle if their vault metadata is inconsistent.
How is this different from asking AI for a .base file?
A generic prompt may produce plausible YAML that does not render. The obsidian-bases skill is better because it is centered on the actual .base workflow: filters first, formulas second, views third, validation last. It also points you to the function reference when formulas are involved.
When is obsidian-bases the wrong tool?
Do not use obsidian-bases if your real problem is missing metadata, inconsistent note structure, or uncertainty about what the base should show. It also will not replace plugin-specific documentation outside the Bases format. If your vault schema is messy, fix that first.
Does obsidian-bases help with formulas and functions?
Yes. The included references/FUNCTIONS_REFERENCE.md is one of the most useful parts of the skill for install-decision purposes. It helps when you need date math, conditional display, type conversion, links, icons, or HTML rendering inside formula-driven properties.
How to Improve obsidian-bases skill
Give cleaner vault schema before asking for output
The biggest improvement lever is input quality. List the exact property names and example values from 3-5 notes before asking for a full base. If one note uses due-date and another uses due, the agent cannot reliably design filters or formulas.
Prevent the most common obsidian-bases failures
Common failure modes:
- invalid YAML from unquoted strings
- formula references used before definition
- filters targeting tags or properties that do not exist
- overcomplicated first draft with multiple views and summaries at once
A better first request is a narrow one: one folder, one filter, one view, one formula.
Iterate from working minimal output
For better obsidian-bases usage, ask for revision in layers:
- first pass: valid
.basewith one filter and one view - second pass: computed fields
- third pass: alternate views and summaries
- final pass: comment on assumptions and possible schema mismatches
This improves reliability because you can see exactly which addition breaks rendering.
Use the reference file strategically
To improve obsidian-bases results, do not ask for “advanced formulas” in the abstract. Name the operation you want and have the agent map it to known functions from references/FUNCTIONS_REFERENCE.md. That keeps the output grounded in the documented function set and reduces hallucinated syntax, especially for dates, booleans, links, and display formatting.