Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Bootstrap a fresh LLM Wiki vault with the three-layer structure, schema files, and starter templates. Usage /wiki-init <path> --topic "<topic>" [--tool all|claude-code|codex|cursor|antigravity]
.claude/skills/alirezarezvani-wiki-init/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -74% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -23% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -42% | 0% |
<!-- canonical copy: engineering/llm-wiki/commands/wiki-init.md — keep in sync -->
Bootstrap a new LLM Wiki vault. Creates raw/, wiki/{entities,concepts,sources,comparisons,synthesis}, the index and log, and installs the schema file(s) for your LLM CLI of choice.
/wiki-init <path> --topic "<one-line topic>"
/wiki-init <path> --topic "<topic>" --tool <claude-code|codex|cursor|antigravity|opencode|gemini-cli|all>
/wiki-init <path> --topic "<topic>" --force # overwrite non-empty dir/wiki-init ~/vaults/research --topic "LLM interpretability"
/wiki-init ./book-wiki --topic "The Power Broker — Robert Caro" --tool all
/wiki-init ~/vaults/founders --topic "SaaS founder playbook" --tool codex<path>/
├── raw/
│ └── assets/
├── wiki/
│ ├── index.md # from template
│ ├── log.md # from template
│ ├── entities/
│ ├── concepts/
│ ├── sources/
│ ├── comparisons/
│ ├── synthesis/
│ └── .templates/ # page templates for reference
├── CLAUDE.md # if --tool claude-code or all
├── AGENTS.md # if --tool codex|cursor|antigravity|opencode|gemini-cli|all
├── .cursorrules # if --tool cursor or all
└── .gitignoreAfter init:
raw//wiki-ingest raw/<your-file>engineering/llm-wiki/skills/llm-wiki/scripts/init_vault.py→ engineering/llm-wiki/skills/llm-wiki/SKILL.md
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 22,176 | 3,041 | -86% | 1 | 1 | 0% | 3,998 | 1,027 | -74% | 0 | 0 | — |
case-02 | fail→fail | 16,813 | 5,634 | -66% | 1 | 1 | 0% | 3,109 | 1,435 | -54% | 0 | 0 | — |
case-03 | fail→fail | 27,854 | 4,923 | -82% | 1 | 1 | 0% | 6,180 | 841 | -86% | 0 | 0 | — |
case-04 | fail→pass | 11,139 | 2,911 | -74% | 1 | 1 | 0% | 1,826 | 1,113 | -39% | 0 | 0 | — |
case-05 | fail→pass | 9,259 | 3,105 | -66% | 1 | 1 | 0% | 1,646 | 1,113 | -32% | 0 | 0 | — |
case-06 | fail→fail | 7,383 | 2,228 | -70% | 1 | 1 | 0% | 1,368 | 922 | -33% | 0 | 0 | — |
case-07 | fail→pass | 7,430 | 3,071 | -59% | 1 | 1 | 0% | 1,450 | 1,121 | -23% | 0 | 0 | — |
case-08 | fail→pass | 8,816 | 2,515 | -71% | 1 | 1 | 0% | 1,619 | 945 | -42% | 0 | 0 | — |
case-09 | fail→pass | 9,357 | 2,847 | -70% | 1 | 1 | 0% | 1,627 | 880 | -46% | 0 | 0 | — |
case-10 | fail→pass | 8,795 | 2,339 | -73% | 1 | 1 | 0% | 1,753 | 936 | -47% | 0 | 0 | — |
case-11 | fail→pass | 21,992 | 2,622 | -88% | 1 | 1 | 0% | 3,966 | 981 | -75% | 0 | 0 | — |
case-12 | fail→pass | 6,711 | 2,014 | -70% | 1 | 1 | 0% | 1,159 | 853 | -26% | 0 | 0 | — |
case-13 | fail→pass | 13,880 | 5,843 | -58% | 1 | 1 | 0% | 2,303 | 1,691 | -27% | 0 | 0 | — |
case-14 | pass→pass | 6,157 | 1,650 | -73% | 1 | 1 | 0% | 1,016 | 777 | -24% | 0 | 0 | — |
case-15 | pass→pass | 12,691 | 2,560 | -80% | 1 | 1 | 0% | 2,114 | 692 | -67% | 0 | 0 | — |
case-16 | fail→pass | 9,417 | 7,843 | -17% | 1 | 1 | 0% | 1,984 | 2,243 | +13% | 0 | 0 | — |
case-17 | fail→pass | 10,026 | 3,373 | -66% | 1 | 1 | 0% | 1,775 | 1,146 | -35% | 0 | 0 | — |
case-18 | fail→pass | 9,737 | 1,914 | -80% | 1 | 1 | 0% | 1,645 | 740 | -55% | 0 | 0 | — |
case-23 | pass→pass | 6,783 | 1,973 | -71% | 1 | 1 | 0% | 1,191 | 883 | -26% | 0 | 0 | — |
case-19 | pass→pass | 3,489 | 1,181 | -66% | 1 | 1 | 0% | 554 | 686 | +24% | 0 | 0 | — |
case-20 | fail→pass | 13,160 | 6,204 | -53% | 1 | 1 | 0% | 2,618 | 1,635 | -38% | 0 | 0 | — |
case-21 | pass→pass | 11,244 | 5,721 | -49% | 1 | 1 | 0% | 2,157 | 1,577 | -27% | 0 | 0 | — |
case-22 | pass→pass | 11,909 | 8,371 | -30% | 1 | 1 | 0% | 2,227 | 2,046 | -8% | 0 | 0 | — |
case-24 | fail→pass | 6,061 | 2,068 | -66% | 1 | 1 | 0% | 1,025 | 837 | -18% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted, and 23 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +63 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.