Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Tactic for compiling research findings into vault pages — orchestrates page creation, updates, edge linking, and index maintenance. Minimum yield ≥3 page operations per invocation.
.claude/skills/yogsoth-ai-knowledge-compilation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-09 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -35% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -40% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 87% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 63% | 0% |
Compile research findings into structured vault pages. Triggered after each strategy completes, or when CC judges accumulated information density warrants compilation.
wiki-search — check for existing pages before creating duplicateswiki-ingest-source — write immutable source pageswiki-compile-page — create/update wiki pages with edgeswiki-add-edge — create typed relationships between pageswiki-lint-fix — run lint in report mode at end<HARD-GATE> ≥3 page operations per invocation (any combination of: create page, update page, add edge). If incoming material cannot produce 3 operations, accumulate and defer. </HARD-GATE>
CC has full autonomy over:
Typical flow (reference, not prescription):
After compilation completes, trigger context-checkpoint to persist vault state changes to the context layer.
<!-- BEGIN available-tables (generated) -->
Optional, no fixed order; the final leaf is always a sop.
| SOP | When to use | | --- | --- | | wiki-add-edge | SOP wrapping vault_add_edge — create a typed relationship between two vault pages. | | wiki-compile-page | SOP for wiki page creation/update — create or update a synthesized wiki page with edges and index update. | | wiki-edge-audit | SOP for auditing wikilink coverage — scans all edges and reports which source pages are missing dir/slug]] wikilinks to their targets. | | wiki-graph-query | SOP wrapping vault_query_graph — traverse knowledge graph from a node to explore its neighborhood and context. | | wiki-ingest-source | SOP for source page creation — write immutable source page capturing raw material, then update search index. | | wiki-lint-fix | SOP wrapping vault_lint — run batch validation, report issues, optionally auto-fix safe problems. | | wiki-search | SOP wrapping vault_search — BM25 full-text search across vault pages. Returns ranked results with snippets. |
<!-- END available-tables (generated) -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 14,641 | 113,791 | +677% | 1 | 1 | 0% | 302 | 1,083 | +259% | 0 | 0 | — |
case-02 | fail→fail | 48,892 | 42,376 | -13% | 1 | 1 | 0% | 8,240 | 1,276 | -85% | 0 | 0 | — |
case-03 | fail→fail | 25,536 | 47,534 | +86% | 1 | 1 | 0% | 1,114 | 1,101 | -1% | 0 | 0 | — |
case-04 | fail→fail | 11,546 | 53,018 | +359% | 1 | 1 | 0% | 915 | 5,229 | +471% | 0 | 0 | — |
case-05 | pass→pass | 13,585 | 18,091 | +33% | 1 | 1 | 0% | 1,428 | 2,333 | +63% | 0 | 0 | — |
case-06 | fail→fail | 25,988 | 51,432 | +98% | 1 | 1 | 0% | 4,724 | 2,599 | -45% | 0 | 0 | — |
case-07 | fail→fail | 15,308 | 39,846 | +160% | 1 | 1 | 0% | 2,458 | 1,104 | -55% | 0 | 0 | — |
case-08 | pass→pass | 23,762 | 27,168 | +14% | 1 | 1 | 0% | 2,070 | 1,657 | -20% | 0 | 0 | — |
case-09 | fail→pass | 11,881 | 4,888 | -59% | 1 | 1 | 0% | 1,893 | 894 | -53% | 0 | 0 | — |
case-10 | fail→pass | 16,163 | 6,537 | -60% | 1 | 1 | 0% | 2,530 | 1,641 | -35% | 0 | 0 | — |
case-19 | pass→pass | 7,171 | 22,891 | +219% | 1 | 1 | 0% | 1,480 | 3,168 | +114% | 0 | 0 | — |
case-11 | pass→pass | 12,997 | 10,324 | -21% | 1 | 1 | 0% | 1,961 | 1,999 | +2% | 0 | 0 | — |
case-12 | fail→pass | 15,578 | 8,057 | -48% | 1 | 1 | 0% | 1,962 | 1,168 | -40% | 0 | 0 | — |
case-13 | fail→fail | 8,418 | 7,610 | -10% | 1 | 1 | 0% | 1,342 | 1,065 | -21% | 0 | 0 | — |
case-14 | pass→pass | 15,588 | 12,810 | -18% | 1 | 1 | 0% | 1,595 | 1,717 | +8% | 0 | 0 | — |
case-15 | pass→pass | 11,614 | 12,212 | +5% | 1 | 1 | 0% | 1,927 | 1,475 | -23% | 0 | 0 | — |
case-16 | fail→pass | 8,453 | 15,487 | +83% | 1 | 1 | 0% | 1,415 | 2,646 | +87% | 0 | 0 | — |
case-17 | pass→pass | 21,330 | 7,795 | -63% | 1 | 1 | 0% | 2,133 | 1,499 | -30% | 0 | 0 | — |
case-18 | pass→pass | 16,190 | 24,379 | +51% | 1 | 1 | 0% | 2,529 | 4,334 | +71% | 0 | 0 | — |
case-20 | pass→pass | 8,408 | 9,049 | +8% | 1 | 1 | 0% | 1,227 | 2,406 | +96% | 0 | 0 | — |
case-21 | pass→pass | 13,515 | 8,548 | -37% | 1 | 1 | 0% | 2,161 | 1,758 | -19% | 0 | 0 | — |
case-22 | pass→pass | 8,963 | 5,670 | -37% | 1 | 1 | 0% | 1,374 | 1,224 | -11% | 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. 22 cases were attempted, and 17 counted toward the lift figure. The other 5 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 +18 percentage points is the difference between those two pass rates over the 17 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.