Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Add gcov code coverage instrumentation to C/C++ projects
.claude/skills/gadievron-code-coverage-with-gcov/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 8 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -5% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 386% | 0% |
| case-06 | ✓→✓ | = Same ✓ | -9% | 0% |
| case-04 | ✓→✓ | = Same ✓ | -2% | 0% |
Instrument C/C++ programs with gcov to measure test coverage.
bashgcc --coverage -o program source.c
bash./program # Creates .gcda files with execution data
Text report:
bashgcov source.c # Creates source.c.gcov with line-by-line coverage
HTML report:
bashgcovr --html-details -o coverage.html
--coverage (shorthand for -fprofile-arcs -ftest-coverage -lgcov)CFLAGS and LDFLAGSmakefileENABLE_COVERAGE ?= 0 ifeq ($(ENABLE_COVERAGE),1) CFLAGS += --coverage LDFLAGS += --coverage endif
cmakeoption(ENABLE_COVERAGE "Enable coverage" OFF) if(ENABLE_COVERAGE) add_compile_options(--coverage) add_link_options(--coverage) endif()
--coverage to CFLAGS and LDFLAGSmake clean or rm -f *.gcda *.gcnomake ENABLE_COVERAGE=1 or cmake -DENABLE_COVERAGE=ONmake test or ./test_suitegcovr --html-details coverage.html --print-summaryText (.gcov files):
-: 0:Source:main.c
5: 42: int x = 10;
#####: 43: unused_code();5: = executed 5 times#####: = not executed-: = non-executableHTML: Interactive report with color-coded coverage
Target: 80%+ line coverage, 70%+ branch coverage
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→pass | 12,633 | 9,110 | -28% | 1 | 1 | 0% | 2,430 | 2,222 | -9% | 0 | 0 | — |
case-01 | fail→pass | 15,145 | 8,535 | -44% | 1 | 1 | 0% | 3,122 | 2,247 | -28% | 0 | 0 | — |
case-02 | fail→pass | 13,545 | 8,829 | -35% | 1 | 1 | 0% | 2,634 | 2,513 | -5% | 0 | 0 | — |
case-03 | fail→pass | 2,296 | 5,714 | +149% | 1 | 1 | 0% | 331 | 1,608 | +386% | 0 | 0 | — |
case-04 | pass→pass | 13,938 | 11,001 | -21% | 1 | 1 | 0% | 2,786 | 2,737 | -2% | 0 | 0 | — |
case-05 | pass→pass | 16,304 | 11,896 | -27% | 1 | 1 | 0% | 3,101 | 2,766 | -11% | 0 | 0 | — |
case-07 | pass→pass | 3,254 | 3,450 | +6% | 1 | 1 | 0% | 574 | 1,184 | +106% | 0 | 0 | — |
case-08 | pass→pass | 7,073 | 3,478 | -51% | 1 | 1 | 0% | 1,306 | 1,157 | -11% | 0 | 0 | — |
case-09 | pass→pass | 9,634 | 4,407 | -54% | 1 | 1 | 0% | 1,733 | 1,279 | -26% | 0 | 0 | — |
case-10 | pass→pass | 7,099 | 3,506 | -51% | 1 | 1 | 0% | 1,375 | 1,165 | -15% | 0 | 0 | — |
case-11 | pass→pass | 3,047 | 2,276 | -25% | 1 | 1 | 0% | 522 | 889 | +70% | 0 | 0 | — |
case-12 | pass→pass | 4,314 | 2,717 | -37% | 1 | 1 | 0% | 686 | 1,001 | +46% | 0 | 0 | — |
case-13 | pass→pass | 9,311 | 5,158 | -45% | 1 | 1 | 0% | 1,705 | 1,495 | -12% | 0 | 0 | — |
case-14 | pass→pass | 14,032 | 7,086 | -50% | 1 | 1 | 0% | 2,285 | 1,781 | -22% | 0 | 0 | — |
case-19 | pass→pass | 6,787 | 2,351 | -65% | 1 | 1 | 0% | 1,256 | 993 | -21% | 0 | 0 | — |
case-15 | pass→pass | 10,069 | 9,759 | -3% | 1 | 1 | 0% | 1,905 | 2,619 | +37% | 0 | 0 | — |
case-16 | pass→pass | 12,998 | 11,810 | -9% | 1 | 1 | 0% | 2,698 | 3,025 | +12% | 0 | 0 | — |
case-17 | pass→pass | 8,708 | 7,619 | -13% | 1 | 1 | 0% | 1,602 | 2,080 | +30% | 0 | 0 | — |
case-18 | pass→pass | 9,473 | 4,241 | -55% | 1 | 1 | 0% | 1,676 | 1,339 | -20% | 0 | 0 | — |
case-20 | pass→pass | 3,947 | 4,072 | +3% | 1 | 1 | 0% | 705 | 1,022 | +45% | 0 | 0 | — |
case-21 | pass→pass | 5,645 | 2,186 | -61% | 1 | 1 | 0% | 887 | 885 | -0% | 0 | 0 | — |
case-22 | pass→pass | 2,942 | 2,237 | -24% | 1 | 1 | 0% | 405 | 890 | +120% | 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. The headline lift of +14 percentage points is the difference between those two pass rates over the 22 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.