coverage-analysis

# Coverage Analysis Agent analyzes .NET coverage data and ranks risky low-coverage code. ## When to Use | User prompt | Use | |---|---| | User asks why coverage stalled | Use this skill | | User asks where tests add the most value | Use this skill | | User asks for coverage hotspots from Cobertura data | Use this skill | ## When Not to Use | User prompt | Route | |---|---| | User asks for one method CRAP score | Use `crap-score` | | User asks to write tests | Use `code-testing-agent` | | User asks for raw HTML only | Generate the short report first | ## Required Inputs | Input | Required | Description | |---|---|---| | Source scope | Yes | Project or solution path | | Cobertura files | No | Reuse existing files first | | Coverage thresholds | No | Use when the user gives pass bands | | Output path | No | Default to `test-results/coverage-analysis/` | ## Workflow | Step | Agent action | Test | Pass | |---|---|---|---| | 1 | Agent reads the scope and looks for Cobertura XML. | Run `rg -n "coverage\.cobertura\.xml|cobertura" <scope>`. | Existing coverage files are listed or none are found. | | 2 | Agent reuses existing coverage files when files exist. | Read the selected file paths. | Agent skips new test runs. | | 3 | Agent generates coverage only when no usable Cobertura file exists. | Run the repo coverage command or `dotnet test` coverage command. | A Cobertura XML file exists under `test-results`. | | 4 | Agent preserves raw inputs under `test-results/coverage-analysis/raw/`. | Read `test-results/coverage-analysis/raw/`. | Raw coverage files exist. | | 5 | Agent computes line coverage, branch coverage, and hotspot ranking. | Read the generated summary. | Summary includes overall coverage and hotspot rows. | | 6 | Agent writes `test-results/coverage-analysis/coverage-analysis.md`. | Read the markdown file. | File exists and contains summary, hotspots, and next steps. | | 7 | Agent generates HTML or CSV only when the user asks. | Read the user prompt. | Extra artifacts exist only for explicit requests. | ## Report Content | Section | Requirement | |---|---| | Coverage summary | Include line coverage and branch coverage | | Hotspots | List risky low-coverage methods or components | | Recommendations | List 1-3 next test targets | | Limits | State blockers or missing data | ## Verification Checklist - [ ] Agent reused existing Cobertura data first. - [ ] Agent ran coverage only when reuse failed. - [ ] Agent preserved raw coverage inputs. - [ ] Agent wrote a short markdown summary. - [ ] Agent ranked hotspots, not percentages only. - [ ] Agent kept optional rich artifacts behind an explicit user request. ## Common Pitfalls | Pitfall | Agent fix | |---|---| | Agent runs expensive tests first | Agent reads for Cobertura files first. | | Agent reports percentages only | Agent adds hotspot ranking and next steps. | | Agent overwrites user files | Agent writes under `test-results/coverage-analysis/`. | | Agent generates HTML by default | Agent writes the markdown summary first. |