Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Fast codebase search via WarpGrep (20x faster than grep)
.claude/skills/morph-search/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 646% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 101% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-14 | ✗→✓ | ▲ Improved | -50% | 0% |
Fast, AI-powered codebase search using WarpGrep. 20x faster than traditional grep.
bashuv run python -m runtime.harness scripts/mcp/morph_search.py \ --search "authentication" --path "."
bashuv run python -m runtime.harness scripts/mcp/morph_search.py \ --search "def.*login" --path "./src"
bashuv run python -m runtime.harness scripts/mcp/morph_search.py \ --edit "/path/to/file.py" --content "new content"
| Parameter | Description | |-----------|-------------| | --search | Search query/pattern | | --path | Directory to search (default: .) | | --edit | File path to edit | | --content | New content for file (use with --edit) |
bash# Find all async functions uv run python -m runtime.harness scripts/mcp/morph_search.py \ --search "async def" --path "./src" # Search for imports uv run python -m runtime.harness scripts/mcp/morph_search.py \ --search "from fastapi import" --path "."
| Tool | Best For | |------|----------| | morph/warpgrep | Fast text/regex search (20x faster) | | ast-grep | Structural code search (understands syntax) |
Requires morph server in mcp_config.json with MORPH_API_KEY.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,575 | 4,428 | -42% | 1 | 1 | 0% | 1,285 | 630 | -51% | 0 | 0 | — |
case-02 | fail→fail | 5,262 | 4,968 | -6% | 1 | 1 | 0% | 857 | 761 | -11% | 0 | 0 | — |
case-03 | fail→pass | 2,198 | 12,900 | +487% | 1 | 1 | 0% | 197 | 1,470 | +646% | 0 | 0 | — |
case-04 | fail→pass | 8,109 | 1,756 | -78% | 1 | 1 | 0% | 1,407 | 691 | -51% | 0 | 0 | — |
case-05 | fail→fail | 9,876 | 7,389 | -25% | 1 | 1 | 0% | 1,145 | 599 | -48% | 0 | 0 | — |
case-06 | fail→pass | 12,319 | 1,988 | -84% | 1 | 1 | 0% | 287 | 578 | +101% | 0 | 0 | — |
case-07 | fail→pass | 4,539 | 1,648 | -64% | 1 | 1 | 0% | 372 | 641 | +72% | 0 | 0 | — |
case-08 | fail→fail | 6,625 | 13,978 | +111% | 1 | 1 | 0% | 945 | 1,992 | +111% | 0 | 0 | — |
case-09 | fail→fail | 5,245 | 25,269 | +382% | 1 | 1 | 0% | 939 | 605 | -36% | 0 | 0 | — |
case-10 | pass→pass | 4,931 | 1,894 | -62% | 1 | 1 | 0% | 912 | 691 | -24% | 0 | 0 | — |
case-11 | pass→pass | 13,543 | 6,421 | -53% | 1 | 1 | 0% | 2,187 | 1,444 | -34% | 0 | 0 | — |
case-12 | fail→fail | 3,158 | 4,077 | +29% | 1 | 1 | 0% | 320 | 588 | +84% | 0 | 0 | — |
case-13 | fail→fail | 2,847 | 11,849 | +316% | 1 | 1 | 0% | 418 | 1,283 | +207% | 0 | 0 | — |
case-14 | fail→pass | 10,290 | 1,192 | -88% | 1 | 1 | 0% | 1,170 | 586 | -50% | 0 | 0 | — |
case-15 | fail→fail | 21,488 | 4,746 | -78% | 1 | 1 | 0% | 2,793 | 607 | -78% | 0 | 0 | — |
case-16 | fail→pass | 4,155 | 3,973 | -4% | 1 | 1 | 0% | 156 | 674 | +332% | 0 | 0 | — |
case-17 | fail→pass | 11,866 | 1,480 | -88% | 1 | 1 | 0% | 2,346 | 601 | -74% | 0 | 0 | — |
case-18 | fail→fail | 4,684 | 11,730 | +150% | 1 | 1 | 0% | 201 | 703 | +250% | 0 | 0 | — |
case-19 | pass→pass | 15,935 | 11,575 | -27% | 1 | 1 | 0% | 2,577 | 2,229 | -14% | 0 | 0 | — |
case-20 | pass→pass | 11,059 | 4,633 | -58% | 1 | 1 | 0% | 2,071 | 1,196 | -42% | 0 | 0 | — |
case-21 | pass→pass | 5,044 | 3,966 | -21% | 1 | 1 | 0% | 806 | 1,148 | +42% | 0 | 0 | — |
case-22 | pass→pass | 6,280 | 5,994 | -5% | 1 | 1 | 0% | 1,006 | 1,460 | +45% | 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 10 counted toward the lift figure. The other 12 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 10 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/31/2026 | +57% |
Other measured skills in the registry, with their headline benchmark lift.