Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Ingest and normalize market data into OHLCV vectors with HNSW indexing
.claude/skills/ruvnet-market-ingest/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 59% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -58% | 0% |
Fetch market data for a symbol, normalize to OHLCV vectors, and store with HNSW indexing for fast pattern search.
When you need to ingest raw market data (price and volume) for a symbol and prepare it for pattern detection and similarity search. This is the first step before running pattern detection or comparison.
(open - prev_close) / prev_close(high - open) / open(low - open) / open(close - open) / openmcp__plugin_ruflo-core_ruflo__embeddings_generate (NOT embeddings_embed — that tool name does not exist).mcp__plugin_ruflo-core_ruflo__memory_store --namespace market-data to persist normalized OHLCV data with symbol+date keys. The memory_* tool family routes by namespace; the agentdb_hierarchical-* family routes by tier (working|episodic|semantic) and ignores namespace strings, so use memory_* here.mcp__plugin_ruflo-core_ruflo__ruvllm_hnsw_add to add vectors to the HNSW index for nearest-neighbor search.bashnpx @claude-flow/cli@latest memory store --namespace market-data --key "symbol-SYMBOL-DATE" --value "OHLCV_JSON"
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-04 | pass→pass | 12,993 | 2,127 | -84% | 1 | 1 | 0% | 2,408 | 827 | -66% | 0 | 0 | — |
case-01 | fail→fail | 12,512 | 41,159 | +229% | 1 | 1 | 0% | 2,403 | 4,408 | +83% | 0 | 0 | — |
case-02 | fail→fail | 11,617 | 20,899 | +80% | 1 | 1 | 0% | 2,521 | 5,676 | +125% | 0 | 0 | — |
case-03 | fail→pass | 10,116 | 12,335 | +22% | 1 | 1 | 0% | 1,946 | 3,089 | +59% | 0 | 0 | — |
case-05 | pass→pass | 9,676 | 2,833 | -71% | 1 | 1 | 0% | 1,988 | 1,030 | -48% | 0 | 0 | — |
case-06 | pass→pass | 15,241 | 3,734 | -76% | 1 | 1 | 0% | 2,669 | 1,045 | -61% | 0 | 0 | — |
case-07 | fail→pass | 11,844 | 3,151 | -73% | 1 | 1 | 0% | 2,237 | 1,103 | -51% | 0 | 0 | — |
case-08 | fail→fail | 5,211 | 1,712 | -67% | 1 | 1 | 0% | 933 | 760 | -19% | 0 | 0 | — |
case-09 | fail→pass | 13,741 | 4,332 | -68% | 1 | 1 | 0% | 2,380 | 1,312 | -45% | 0 | 0 | — |
case-10 | fail→pass | 11,473 | 3,693 | -68% | 1 | 1 | 0% | 2,201 | 1,332 | -39% | 0 | 0 | — |
case-11 | fail→fail | 5,257 | 1,582 | -70% | 1 | 1 | 0% | 1,023 | 765 | -25% | 0 | 0 | — |
case-12 | fail→fail | 10,900 | 1,651 | -85% | 1 | 1 | 0% | 2,152 | 743 | -65% | 0 | 0 | — |
case-13 | fail→pass | 14,166 | 2,815 | -80% | 1 | 1 | 0% | 2,472 | 1,031 | -58% | 0 | 0 | — |
case-14 | pass→pass | 12,410 | 1,688 | -86% | 1 | 1 | 0% | 2,202 | 728 | -67% | 0 | 0 | — |
case-15 | fail→fail | 11,947 | 2,404 | -80% | 1 | 1 | 0% | 2,224 | 964 | -57% | 0 | 0 | — |
case-16 | fail→pass | 12,728 | 3,871 | -70% | 1 | 1 | 0% | 2,202 | 1,121 | -49% | 0 | 0 | — |
case-17 | fail→pass | 6,108 | 2,404 | -61% | 1 | 1 | 0% | 1,204 | 996 | -17% | 0 | 0 | — |
case-18 | fail→pass | 12,459 | 2,077 | -83% | 1 | 1 | 0% | 2,093 | 820 | -61% | 0 | 0 | — |
case-19 | pass→pass | 12,498 | 3,415 | -73% | 1 | 1 | 0% | 2,229 | 1,159 | -48% | 0 | 0 | — |
case-20 | fail→fail | 7,471 | 7,319 | -2% | 1 | 1 | 0% | 1,426 | 1,307 | -8% | 0 | 0 | — |
case-21 | fail→fail | 4,442 | 13,594 | +206% | 1 | 1 | 0% | 868 | 2,502 | +188% | 0 | 0 | — |
case-22 | pass→fail | 5,588 | 8,657 | +55% | 1 | 1 | 0% | 1,001 | 1,093 | +9% | 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 19 counted toward the lift figure. The other 3 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 +32 percentage points is the difference between those two pass rates over the 19 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.
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.