A fast, cross-platform CLI tool that parses JaCoCo XML coverage reports, filters out trivial and fully-covered methods, and scores the remaining methods by complexity × coverage gap. The compact JSON output is designed to be fed into Claude Code to guide targeted test generation.

• Filters constructors, getters, and setters automatically
• Skips fully-covered methods (no missed lines)
• Scores each method: complexity × (missed_lines / total_lines)
• Reports overall and per-class line coverage percentage (--summary)
• Outputs compact or pretty-printed JSON
• Supports a minimum score threshold to focus on high-priority gaps
• Single static binary — no runtime dependencies

git clone <repo-url>
cd jacoco-filter
cargo install --path .

The jacoco-filter binary will be installed to ~/.cargo/bin/ and available on your PATH. Alternatively, build without installing:

cargo build --release
# Binary at target/release/jacoco-filter

mvn clean test jacoco:report
# Report is written to target/site/jacoco/jacoco.xml

# Print to stdout (compact JSON — methods array)
jacoco-filter target/site/jacoco/jacoco.xml

# Include overall line coverage summary
jacoco-filter target/site/jacoco/jacoco.xml --summary --pretty

# Only show methods with a score above 1.5
jacoco-filter target/site/jacoco/jacoco.xml --min-score 1.5 --pretty

# Pretty-print and save to a file
jacoco-filter target/site/jacoco/jacoco.xml --summary --pretty --output coverage-gaps.json

If you use Claude Code in your Java project, copy the bundled skill so Claude knows how to run jacoco-filter automatically:

mkdir -p /path/to/your-java-project/.claude/skills
cp -r skill/jacoco-filter /path/to/your-java-project/.claude/skills/

Once installed, Claude Code will automatically use jacoco-filter whenever you ask it to analyze coverage — and will never attempt to read jacoco.xml directly.

claude "Here are the highest-priority coverage gaps. Please write tests for these methods." \
  --file coverage-gaps.json

Results are sorted by score descending:

[
  {
    "class": "com.example.billing.InvoiceService",
    "source_file": "InvoiceService.java",
    "method": "calculateDiscount",
    "score": 4.5,
    "missed_lines": [42, 43, 47, 51]
  }
]

Wraps the methods array in an object with a coverage summary:

{
  "summary": {
    "line_coverage_pct": 72.4,
    "lines_covered": 842,
    "lines_missed": 321,
    "by_class": [
      {
        "class": "com.example.billing.InvoiceService",
        "source_file": "InvoiceService.java",
        "line_coverage_pct": 45.0,
        "lines_covered": 9,
        "lines_missed": 11
      }
    ]
  },
  "methods": [ ... ]
}

by_class is sorted ascending by line_coverage_pct (worst-covered first). Use jq to pull out just the number you need:

# Overall coverage percentage
jacoco-filter jacoco.xml --summary | jq '.summary.line_coverage_pct'

# Pass/fail against an 80% target
jacoco-filter jacoco.xml --summary | jq 'if .summary.line_coverage_pct >= 80 then "PASS" else "FAIL" end'

# Five worst-covered classes
jacoco-filter jacoco.xml --summary | jq '.summary.by_class[:5] | [.[] | {class, line_coverage_pct}]'

jacoco-filter <input_file> [OPTIONS]

Arguments:
  <input_file>   Path to jacoco.xml

Options:
  --output <path>       Write JSON to file (default: stdout)
  --min-score <float>   Exclude methods below this score (default: 0.0)
  --pretty              Pretty-print JSON output
  --summary             Include line-coverage summary in output
  -h, --help            Print help
  -V, --version         Print version

score = complexity × (missed_line_count / total_line_count)

• complexity — sum of missed + covered from the COMPLEXITY counter
• missed_line_count — number of lines attributed to this method where mi > 0 or mb > 0
• total_line_count — sum of missed + covered from the LINE counter

A score of 0.0 means either the method is fully covered or has no lines.

cargo build        # debug build
cargo test         # run all tests
cargo clippy       # lint
cargo fmt          # format

MIT