Skip to content
Back to skills

Atlas Grow

ASecurity

The next step. Propose, with evidence, what would add the most to this vault. Writes nothing until a person picks. Use for "what should I fill in next" · "find the gaps" · "what's missing".

  • 141 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 2, 2026
developmentgonode

Security analysis

A100/100

Scanned September 2, 2026

npx -y skills add wlsdks/ontology-atlas --skill atlas-grow --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Atlas Grow?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Atlas Grow
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/wlsdks-atlas-grow-ontology-atlas/badge)](https://www.skillsdirectory.com/skills/wlsdks-atlas-grow-ontology-atlas)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: atlas-grow
description: The next step. Propose, with evidence, what would add the most to this vault. Writes nothing until a person picks. Use for "what should I fill in next" · "find the gaps" · "what's missing".
---

# /atlas-grow — where to fill in next

## One rule — the proposal step never writes

Only what a person picked lands in the vault. An invented node is the worst
thing this tool can produce: **a wrong map is worse than no map.** With no map
people go look for themselves; with a wrong one they decide on it.

1. **Where does this vault stand**

- `query_ontology({ operation: 'agent_brief' })` — the starting point and next action
- `query_ontology({ operation: 'growth_plan' })` — what is empty

2. **Filter candidates against evidence**

Do not relay the candidates as they arrive. Each one must survive three tests.

- **Does it exist** — is there evidence in code or docs, or does it merely
   sound plausible? Check with `find_evidence({ title })`. If a source folder is
   bound, `analyze_repo_structure` and `infer_imports` supply the evidence.
- **Is it already here** — search near-names first with
   `query_concepts({ filter })`. If it exists, do not create a second one; fill
   the existing node with `patch_concept` (pass `expected_mtime`).
- **Does it carry meaning** — an element needs a reason beyond its location.
   One node per file is not a map, it is a file listing.

3. **Show it to the person**

**Five at a time, at most.** Each line carries three things — ① what ② where it
attaches ③ why you believe it (the evidence). Do not ask "shall I create them
all?" Number them so the answer can be a selection.

4. **Write only what was approved — then verify**

- Nodes: `add_concepts` (chunks of 50 when there are many)
- **Only after every node succeeded**, relations: `add_relations`
- `validate_vault({})` → `query_ontology({ operation: 'health' })`

**A new domain does not attach itself to the project.** Skip
`add_relation('<project>', 'domains/<new domain>', 'domains')` and you have
built something the map will never show. Step 4's health check catches it.

## How this skill fails

- Invents plausible names with no evidence
- Misses an existing node and creates a near-duplicate beside it
- Writes without asking
- Creates nodes but no relations, so nothing it made is visible anywhere
- Skips verification, leaving the person to discover the breakage later

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.