The 22 skills · Craft skill · Core Pack and Everything
AGENT SKILL
workspace-architecture: what it makes your agent do, and what it refuses.
workspace-architecture designs the third surface. An operator works through the agent they talk to, the site it builds, and the folder on their own machine, and the third one is almost always an accident: a Downloads folder and a hope. This skill makes it a layout with rules.
The job workspace-architecture does
The skill runs the first hour in order: survey what is already on disk, agree the layout, then set the boundaries that decide what breaks when you cross them. Raw exports land in one zone, arranged per feed rather than per file type, and stay immutable there. Processed data is regenerable and page agnostic. Configuration holds every tenant fact, so nothing tenant specific leaks into code. Names are addresses rather than assertions, which is why registration beats detection and a receipt ledger records one row per arrival. It also produces the build roadmap on day one, the path from raw files to a working intranet, with every step only a human can decide marked and waiting.
What workspace-architecture refuses
A skill is worth what it stops as much as what it produces. These are the refusals written into workspace-architecture, taken from the rules and limits in its own file.
- ✓Trusting a filename. A name may propose where a file belongs, and the header and the bytes decide.
- ✓Moving the operator's originals. The intake loop copies, and never writes to, renames, or deletes anything in the folder it watches.
- ✓Data outside the data boundary. One ignore boundary, one landing zone, and no source export copied into a report directory.
- ✓A step nobody can run again. Anything done by hand once becomes an entry point in the repository, or the refresh cycle quietly depends on a person.
10 lines from the workspace-architecture skill file, verbatim
From skills/workspace-architecture/SKILL.md, unedited. Kymira skills are plain text you can
read in full before you install them, which is the point of shipping a method rather than a binary.
<project>/ ├─ refresh.py one entry point, every cadence ├─ manifest.json machine-maintained state (never hand-edited) ├─ README.md ARCHITECTURE.md RUNBOOK.md docs stay flat at the root ├─ config/ every tenant fact. Nothing tenant-specific outside it. │ ├─ sources.toml the registered feeds: fingerprints, globs, routing │ ├─ validation.json gate thresholds (configuration, per principle 8) │ ├─ rulings/ definition rulings, attributed and dated │ └─ brand/ logos. Published on every build: nothing else goes here. ├─ data/ ALL data. One gitignore boundary, not many.
The failure modes workspace-architecture addresses
These entries in the catalogue of 51 observed failure modes name this skill as what addresses them. Every one was seen in a real build or in our own audits, and none was invented.
The tier workspace-architecture ships in
| Tier | Price | What the tier holds | workspace-architecture |
|---|---|---|---|
| Free | $0 | honest-dataviz and the eleven doctrine principles, published in full | Not in this tier |
| Core Pack | $199 | the thirteen craft skills, one-time purchase, lifetime updates | Included |
| Everything | $349 | all 22 skills, both libraries, the template bank, one-time purchase, lifetime updates | Included |
Both paid tiers are a one-time purchase with lifetime updates, and all sales are final: the files are plain text you keep the moment they arrive, so judge the quality on the free skill and the published doctrine first. The pricing section states exactly what each tier contains.
Installing workspace-architecture
A skill is a folder with a SKILL.md file in it. Installing one is copying that folder where
your agent looks for skills, with no build step and no dependencies.
mkdir -p ~/.claude/skills cp -R Kymira/skills/workspace-architecture ~/.claude/skills/
Or open the unzipped pack in Claude Code and say install Kymira, which installs every skill in your tier, or double-click the installer for your system. Cursor reads the same skill as a rule file: copy workspace-architecture.mdc into .cursor/rules/. Skills are read at startup, so start a new session and ask your agent which skills it has available.
Not for
A beautifully organised workspace full of wrong figures is still wrong, and this skill cannot tell you whether the numbers inside a well placed file are right. Registration proves identity rather than authority: naming the system of record stays a human ruling, and an existing folder still has to be migrated one feed at a time.
Last updated 2026-08-30 · Dylan, founder · one of 22 packaged skills, each one a method written down, not a product tour.