Full inventory of the master-of skill-gate system — every gated component by domain, plus what is broken and token savings. Invoke when the user asks for the whole list ('전체 보여줘', '다 보여줘', '전체 목록'); for a quick status use check-skills.
Installs into .claude/skills of the current project.
Are you the author of Check Skills All?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/devrudals-check-skills-all)
---
name: "check-skills-all"
description: "Full inventory of the master-of skill-gate system — every gated component by domain, plus what is broken and token savings. Invoke when the user asks for the whole list ('전체 보여줘', '다 보여줘', '전체 목록'); for a quick status use check-skills."
allowed-tools: Read, Bash, Artifact, Skill
---
# check-skills-all
The engine is the `mo` CLI at `~/.master-of/mo`. For a compact overview use
`check-skills` instead — this skill is specifically for the complete,
uncompressed, per-skill list.
`~/.master-of/mo sync` / `claude-sync` do **not** regenerate the human-readable
report — only `mo full` (compact) or `mo full --raw` (complete) does, by
calling the reporter fresh. Reading `~/.master-of/report.txt` directly risks
showing a stale snapshot from whenever it was last rendered; always regenerate
it through `mo full --raw`, never read the file straight off disk.
Because the complete list is long (500+ lines), never retype it into chat —
that costs input tokens to read it AND output tokens to reproduce it, which
routinely exceeds the token savings the report itself is measuring. Instead,
render it once as a small HTML shell and let the browser display the content
directly, so only the shell (a few KB) is generated by you — the report text
itself moves shell-to-file without passing through model output tokens.
1. Regenerate fresh and capture it, without reading it into your own context
yet (silent shell step, no chat output):
```bash
~/.master-of/mo full --raw > "$SCRATCHPAD/report.md"
cp "~/.claude/skills/master-of/skills/check-skills-all/assets/gate-ledger.html" "$SCRATCHPAD/gate-ledger.html"
```
(`$SCRATCHPAD` = your scratchpad directory for this session.)
2. Read `$SCRATCHPAD/report.md` once with the Read tool — required before
publishing any file you did not write yourself with the Write tool, per
the Artifact tool's own rules. This is the only cost that scales with
report size; do not also paste it into your chat reply.
3. Publish via the Artifact tool:
- `file_path`: `$SCRATCHPAD/gate-ledger.html`
- `files`: `{"report.md": "$SCRATCHPAD/report.md"}`
- `favicon`: 🗂️ · `icon`: list
- `description`: one sentence noting this is the live master-of gate inventory.
4. Reply with the published link plus the token-savings figures from the
report's own footer (`토큰 절약 추정치`): before, after, amount saved, and
percentage — e.g. "37,598 tok → 4,951 tok (32,647 saved, 87%)". This is
the only thing the chat reply adds beyond the link; do not restate
per-gate counts, always-on counts, or any of the raw list — that detail
lives in the artifact. If the user is on a surface without artifact
support, fall back to telling them the file lives at
`~/.master-of/report.txt` (after step 1 has refreshed it) so they can open
it directly instead of pasting it into chat.
For a single gate's detail: `~/.master-of/mo gate <name>` (e.g. `mo gate design`).