Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Codebase organization for pikru. Use when you need to find where specific functionality lives.
.claude/skills/bearcove-codebase-layout/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 790% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -39% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -23% | 0% |
src/ - Main pikru library (parsing and rendering)crates/ - Supporting cratesvendor/pikchr-c/ - Original C implementation and test filesexamples/ - Example programstests/ - Integration testssrc/)src/lib.rs - Public API, exports key types and functionssrc/parse.rs - Parser implementation (statement parsing, token handling)src/ast.rs - AST types (Statement, Expr, ObjectClass, etc.)src/types.rs - Core types (ClassName, Direction, EdgePoint, etc.)src/render/mod.rs - Main renderer logicsrc/render/eval.rs - Expression evaluation (positions, scalars, variables)src/render/svg.rs - SVG generation (converts shapes to SVG elements)src/render/geometry.rs - Shape geometry (boxes, circles, paths, files)src/render/shapes.rs - Shape rendering (specific shape implementations)src/render/types.rs - Render types (PObject, Style, PositionedText, etc.)src/render/path_builder.rs - Path construction (line/arrow building)src/render/context.rs - Render context (variable scopes, state)src/render/defaults.rs - Default values (line width, font size, etc.)src/errors.rs - Error types and reportingsrc/macros.rs - Helper macroscrates/pikru-compare/SVG comparison utilities for testing
src/lib.rs - Core comparison logiccompare_outputs() - Main comparison functionCompareResult::is_match() - Determines if test passescrates/pikru-mcp/MCP server for test running
src/tools.rs - MCP tool implementationsrun_pikru_test() - Run single test, returns comparisonlist_pikru_tests() - List available testsdebug_pikru_test() - Run with trace outputsrc/main.rs - MCP server entry pointvendor/pikchr-c/tests/*.pikchr - Test input files from C implementation
test01-test81 - Numbered feature testsautochop*.pikchr - Arrow chopping testssrc/render/mod.rs:1265-1550 - ljust/rjust calculationsrc/render/svg.rs:294 - Text anchor assignmentsrc/render/types.rs - PositionedText structuresrc/render/shapes.rs - Shape-specific renderingsrc/render/geometry.rs - Geometric calculationssrc/render/svg.rs - SVG element generationsrc/parse.rs - Parser logicsrc/ast.rs - AST node definitionscrates/pikru-compare/src/lib.rs - Comparison logicMany functions include // cref: comments pointing to the original C implementation:
// cref: function_name (pikchr.c:line_number)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 41,415 | 4,200 | -90% | 1 | 1 | 0% | 3,059 | 1,721 | -44% | 0 | 0 | — |
case-02 | fail→pass | 6,951 | 12,086 | +74% | 1 | 1 | 0% | 303 | 2,696 | +790% | 0 | 0 | — |
case-03 | fail→pass | 49,699 | 2,886 | -94% | 1 | 1 | 0% | 3,032 | 1,485 | -51% | 0 | 0 | — |
case-04 | fail→pass | 14,433 | 3,101 | -79% | 1 | 1 | 0% | 2,287 | 1,404 | -39% | 0 | 0 | — |
case-05 | fail→pass | 17,767 | 6,431 | -64% | 1 | 1 | 0% | 2,444 | 1,875 | -23% | 0 | 0 | — |
case-06 | fail→pass | 12,297 | 4,168 | -66% | 1 | 1 | 0% | 2,364 | 1,653 | -30% | 0 | 0 | — |
case-07 | fail→pass | 12,049 | 5,746 | -52% | 1 | 1 | 0% | 1,938 | 1,638 | -15% | 0 | 0 | — |
case-08 | fail→pass | 15,367 | 3,224 | -79% | 1 | 1 | 0% | 1,829 | 1,433 | -22% | 0 | 0 | — |
case-09 | pass→pass | 12,899 | 3,925 | -70% | 1 | 1 | 0% | 2,332 | 1,450 | -38% | 0 | 0 | — |
case-10 | fail→pass | 35,210 | 3,258 | -91% | 1 | 1 | 0% | 5,136 | 1,380 | -73% | 0 | 0 | — |
case-11 | fail→pass | 18,158 | 4,286 | -76% | 1 | 1 | 0% | 2,576 | 1,559 | -39% | 0 | 0 | — |
case-12 | fail→pass | 12,764 | 2,657 | -79% | 1 | 1 | 0% | 1,730 | 1,367 | -21% | 0 | 0 | — |
case-13 | fail→pass | 13,322 | 5,229 | -61% | 1 | 1 | 0% | 1,982 | 1,657 | -16% | 0 | 0 | — |
case-14 | pass→pass | 14,063 | 3,017 | -79% | 1 | 1 | 0% | 1,962 | 1,271 | -35% | 0 | 0 | — |
case-15 | fail→pass | 41,245 | 3,380 | -92% | 1 | 1 | 0% | 7,227 | 1,364 | -81% | 0 | 0 | — |
case-16 | fail→pass | 7,663 | 2,959 | -61% | 1 | 1 | 0% | 1,224 | 1,168 | -5% | 0 | 0 | — |
case-17 | pass→pass | 6,005 | 1,562 | -74% | 1 | 1 | 0% | 683 | 1,108 | +62% | 0 | 0 | — |
case-18 | fail→pass | 19,971 | 6,838 | -66% | 1 | 1 | 0% | 2,457 | 2,179 | -11% | 0 | 0 | — |
case-19 | fail→pass | 15,444 | 7,054 | -54% | 1 | 1 | 0% | 2,534 | 1,965 | -22% | 0 | 0 | — |
case-20 | pass→pass | 19,640 | 27,067 | +38% | 1 | 1 | 0% | 3,785 | 5,054 | +34% | 0 | 0 | — |
case-21 | pass→pass | 20,013 | 24,320 | +22% | 1 | 1 | 0% | 2,924 | 4,155 | +42% | 0 | 0 | — |
case-22 | pass→pass | 12,366 | 11,099 | -10% | 1 | 1 | 0% | 2,350 | 2,981 | +27% | 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 21 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 +73 percentage points is the difference between those two pass rates over the 21 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.