Unit Test Coverage Improvement

Unit Test Coverage Improvement

Improve unit test coverage for existing code. Requires /optimus:init to have set up test infrastructure (framework, coverage tooling, testing docs) first. Conservative by design — only adds new test files, never refactors or restructures existing source code. If code is untestable as-is, it flags it rather than changing it. Refactoring is the domain of /optimus:refactor.

Step 1: Pre-flight

Parse invocation arguments

Extract from the user's arguments:

1. deep flag (present/absent)
2. harness keyword after deep (present/absent)
3. A number immediately after deep or deep harness → iteration cap (optional, default 5, hard cap 10)
4. Everything else → scope instructions (optional path)

Examples:

• /optimus:unit-test → normal mode, full project
• /optimus:unit-test src/api → normal mode, scoped
• /optimus:unit-test deep → deep mode (5 iterations)
• /optimus:unit-test deep 3 → deep mode (3 iterations)
• /optimus:unit-test deep src/api → deep mode, scoped
• /optimus:unit-test deep harness → harness mode, 5 iterations
• /optimus:unit-test deep harness 5 "src/api" → harness mode, scoped

Harness mode detection

If system prompt contains HARNESS_MODE_ACTIVE: → Read $CLAUDE_PLUGIN_ROOT/references/coverage-harness-mode.md → Follow the "Unit-Test Phase Execution" section → Skip user confirmation (no AskUserQuestion) → Skip loops — run Steps 2–4 exactly once, then output structured JSON and stop

If harness keyword detected in arguments: → Read $CLAUDE_PLUGIN_ROOT/references/coverage-harness-mode.md Skill-Triggered Invocation section → Pass: max_cycles (from step 1 parsing), scope (from step 1 parsing) → The invocation section will resolve the Python environment, build the terminal command pointing to scripts/test-coverage-harness/main.py, present it, and stop → Do NOT proceed to Step 2

Pre-flight checks

Read $CLAUDE_PLUGIN_ROOT/skills/init/references/multi-repo-detection.md for workspace detection. If a multi-repo workspace is detected, process each repo independently: run Steps 1-5 inside each repo that has .claude/CLAUDE.md. Report results per repo. If no repos have been initialized, suggest running /optimus:init first from the workspace root.

Check that .claude/CLAUDE.md exists. If it doesn't, stop and recommend running /optimus:init first — the project needs baseline context before test generation can be effective.

Beyond the init check, identify which guideline documents are available — they directly affect the quality of everything this skill does:

Document	Role	Effect on skill
coding-guidelines.md	Primary quality reference	Tests follow naming conventions, code structure, quality standards
testing.md	Testing conventions	Framework, runner commands, mocking patterns, file organization
.claude/CLAUDE.md	Project overview	Tech stack signals, test runner commands

Monorepo path note: Read the "Monorepo Scoping Rule" section of $CLAUDE_PLUGIN_ROOT/skills/init/references/constraint-doc-loading.md for doc layout and scoping rules. When generating tests for a subproject, load that subproject's testing.md, not another subproject's.

The skill operates differently depending on what exists:

• All three docs — matches existing conventions precisely
• CLAUDE.md + coding-guidelines (no testing.md) — derives conventions from the codebase
• CLAUDE.md only (no guidelines either) — still works, but with less project-specific guidance

Scope

Parse optional path argument (e.g., /optimus:unit-test src/api) to limit scope. If no path is specified, default to the full project.

For monorepos and multi-repo workspaces, detect project structure using the same approach as /optimus:init — reference $CLAUDE_PLUGIN_ROOT/skills/init/SKILL.md Step 1 for detection logic (multi-repo workspace detection, workspace configs, manifest scanning, supporting signals). Process each project/repo independently.

Interactive deep mode activation

If the deep flag was detected in argument parsing (and harness keyword was NOT), activate deep mode. Deep mode loops discovery-and-write cycles (Steps 2–4) until tests converge or the iteration cap is reached.

Clamp the parsed iteration cap: if it exceeds 10, clamp to 10 and warn: "Iteration cap clamped to 10 (maximum)." If it is less than 1, clamp to 1 and warn: "Iteration cap clamped to 1 (minimum)." Default is 5 when no number is given.

Before proceeding, check whether a test command is available (from .claude/CLAUDE.md). If no test command exists, deep mode's auto-approve loop has no safety net — fall back to normal mode and warn: "Deep mode requires a test command for safe auto-approve. Falling back to normal mode — re-run /optimus:init to set up test infrastructure first." Set deep-mode to false. Then continue with the standard single-pass flow.

If a test command is available, warn the user:

Deep mode runs up to [cap] iterative test-generation passes. Each iteration is a full discovery-and-write cycle — credit and time consumption multiplies with iteration count. Tests are generated automatically at each iteration without per-item approval. Untestable code that requires refactoring is flagged but not addressed; consider running /optimus:refactor testability first or /optimus:unit-test deep harness for an automated unit-test + refactor-testability loop. Each iteration also accumulates context — on large codebases, output quality may degrade in later iterations.

Test command: [test command from CLAUDE.md]

Then use AskUserQuestion — header "Deep mode", question "Proceed with deep mode?":

• Start deep mode — "Run iterative test-generation until converged (max [cap] iterations)"
• Normal mode — "Single pass with manual plan approval instead"

Tell the user: Tip: For large codebases or extended sessions, re-run with /optimus:unit-test deep harness to launch the external harness with fresh context per phase.

If the user did not invoke with deep, skip this step.

If the user selects Normal mode, set deep-mode to false and continue with the standard single-pass flow.

If deep mode is confirmed, set deep-mode to true and initialize:

• Counters: iteration-count = 1, total-added = 0, total-reverted = 0, accumulated-coverage-delta = 0
• accumulated-items = empty list — each entry records file (test file path), target (the file:function/class being tested — one entry per covered symbol, so convergence matching is exact-match on this string), iteration (which iteration added it), and status (pass | reverted — test failure | bug-found | abandoned)
• accumulated-untestable = empty list — items the analyzer classifies as not-testable-without-refactoring across iterations
• accumulated-bugs = empty list — bugs discovered by test failures across iterations

Define the [coverage-summary] placeholder once: when the coverage tool is available, render as <sign><abs(accumulated-coverage-delta)>pp where <sign> is + for values ≥ 0 and - for negative (e.g. +2.5pp, -1.0pp). When the coverage tool is unavailable, render as not measured. Every downstream use (iteration-context block, iteration report, progress summary, cumulative report) applies this same rendering.

Step 2: Discovery & Coverage Analysis (agent-assisted)

Delegate test infrastructure scanning, test execution, and coverage analysis to a reconnaissance agent to keep the main context clean for test writing.

For each subproject (or the single project):

Read $CLAUDE_PLUGIN_ROOT/skills/unit-test/agents/shared-constraints.md for agent constraints. Read $CLAUDE_PLUGIN_ROOT/skills/unit-test/agents/test-infrastructure-analyzer.md for the full prompt template, scanning patterns, execution rules, and return format for the Test Infrastructure Analyzer Agent.

Iteration context injection (interactive deep mode, iterations 2+)

If interactive deep mode is active and iteration-count > 1, prepend a concise context block to the agent prompt before the main instructions. Include:

• Tests already added — bullet list of file → target entries from accumulated-items with status pass, so the agent skips re-discovering the same targets.
• Items previously reverted, abandoned, or bug-found — bullet list from accumulated-items with status reverted — test failure, abandoned, or bug-found, so the agent does not re-propose them.
• Untestable code already flagged — bullet list from accumulated-untestable, so the agent does not re-flag them; focus new discovery on genuinely new items.
• Cumulative coverage delta — [coverage-summary] so far, so the agent has a sense of progress.

The goal is convergence: each iteration should propose new testable items, not duplicates. Keep the block under ~30 lines to limit context drift.

Launch

Launch 1 general-purpose Agent tool call using the prompt from test-infrastructure-analyzer.md, prepended with the shared constraints (and the iteration context block, if applicable).

Agent	Role	Runs when
1 — Test Infrastructure Analysis	Scan test files/frameworks/runners, run existing tests, measure coverage, classify code testability	Always

Wait for the agent to complete.

In interactive deep mode: after the agent returns, append its newly-classified not-testable-without-refactoring entries (from the Testability Classification → Untestable section) to accumulated-untestable, skipping any already present from prior iterations.

Stop gates (evaluated from agent results)

If no test framework is detected in the agent's Discovery Results, stop and report: "No test framework found. Run /optimus:init (or re-run it) to install a test framework and set up test infrastructure before using this skill." Do not proceed to test generation without a working framework.

If the agent's Test Suite Execution reports failures, stop. This skill does not fix failing tests or build-level issues by design. Print the matching message below (Conversation / Mode / Next skill per $CLAUDE_PLUGIN_ROOT/references/skill-handoff.md).

• Fail - assertion (tests compile and run, but some fail): print the quote below, then append a ### Bugs Discovered section listing each failing test as [test file] — [test name] — [one-line failure excerpt] (prefix entries with repo name/path in multi-repo workspaces; omit the excerpt if the test runner output did not expose it).

  Pre-existing tests are failing. A green baseline is required before adding new tests, and this skill does not modify existing tests or source code.

  Next: stay in this conversation (normal mode) and ask Claude to triage the failing tests listed in Bugs Discovered. Once the baseline is green, start a fresh conversation and re-run /optimus:unit-test.

  Tip: for best results, start a fresh conversation for the next skill — each skill gathers its own context from scratch.

• Fail - build (build/bootstrap failures): print the quote below and stop — no Bugs Discovered section (there are no per-test failures to list, only build errors the analyzer has already summarized).

  The test runner cannot start or test files fail to compile. These are build-level issues, not test logic, and /optimus:init owns that repair path.

  Next: start a fresh conversation in normal mode and run /optimus:init — its health check will propose minimal fixes and re-run the suite. Once the build is healthy, start another fresh conversation and re-run /optimus:unit-test.

  Tip: for best results, start a fresh conversation for the next skill — each skill gathers its own context from scratch.

Present to user

From the agent's results, present the Discovery Summary and Coverage Analysis to the user. This sets clear expectations and reinforces the conservative constraint.

Data carried forward

• Test runner command and framework identity → Step 4 (test execution and idioms)
• Coverage tooling → Step 5 (final measurement)
• Testability classification → Step 3 (plan prioritization — testable items become candidates, untestable items are skipped)
• Baseline coverage → Step 5 (before/after comparison)

Step 3: Test Generation Plan

Create a prioritized list, capped at 10 items per run:

1. Exported/public functions and classes — API surface, highest value
2. Pure functions and utility modules — easiest to test, highest ROI
3. Business logic with clear inputs/outputs — core functionality
4. Complex branching logic — high cyclomatic complexity, most likely to have bugs
5. Internal/private helpers — lower priority, test through public API when possible

Skip (flag in summary, don't attempt):

• Code that's untestable without refactoring
• Generated code (protobuf, OpenAPI, ORM migrations)
• Migration files
• Declarative configuration
• Thin wrappers with no logic

User confirmation

Deep mode (harness or interactive): Skip the question — auto-select "Generate tests for all planned items" and proceed directly to Step 4. In interactive deep mode, the deep-mode loop sub-section at the end of Step 4 decides whether to iterate or finalize; in harness mode, Step 6 emits structured JSON and the external harness drives any further cycles.

Normal mode: Present the plan, then use AskUserQuestion — header "Plan", question "How would you like to proceed with the test generation plan?":

• Approve all — "Generate tests for all planned items"
• Selective — "Choose specific items by number"
• Skip — "No tests — keep the plan as reference"

If the user selects Selective, ask which item numbers to proceed with (e.g., "1, 3, 5").

Step 4: Test Writing

Quality standards

Tests must follow:

• coding-guidelines.md for quality standards (naming, structure, clarity)
• testing.md for testing conventions (framework idioms, file naming, directory structure)
• $CLAUDE_PLUGIN_ROOT/skills/tdd/references/testing-anti-patterns.md for mocking discipline — prefer real code over mocks, never assert on mock behavior, mock only external services or non-deterministic dependencies
• Existing test files for concrete patterns to replicate: import style, assertion library, file naming convention, directory placement, shared fixtures (conftest/setup files, factories, beforeAll/setUp blocks), and describe/it or class/method test organization. Extract these patterns before writing the first test and apply them consistently to all generated tests.

Before writing each test

Answer these gate questions — fix any "no" before proceeding:

1. File placement — Does a test file for this module already exist? If yes, add tests there instead of creating a new file. New files must follow the naming convention from testing.md (typically test_<module_name> or <module_name>.test).
2. Fixtures and helpers — Do existing test files or shared setup files (conftest.py, test helpers, factories) already provide fixtures for the data this test needs? Use them instead of creating private helpers that duplicate existing setup.
3. Mocking and assertion discipline — apply the gate questions from $CLAUDE_PLUGIN_ROOT/skills/tdd/references/testing-anti-patterns.md (already referenced in Quality standards above).
4. Setup duplication — apply the DRY principle from coding-guidelines.md to test setup: repeated setup should be extracted to shared fixtures (setUp/beforeEach, conftest, factories).

Conservative constraint

Only add new test files. Never refactor or modify existing source code — refactoring is the domain of /optimus:refactor. If a function can't be tested without changing its signature or extracting dependencies, flag it in the summary instead of changing it.

Per-test workflow

For each approved item:

1. Write the test file
2. Self-review against the "Before writing each test" checklist above — fix any violations before running
3. Run it immediately
4. If the test fails:
   • Fix the test (not the source code) — max 3 fix attempts
   • If still failing after 3 attempts, flag as untestable and move on
   • If the failure reveals an actual bug in existing code, report the bug but do not fix it
5. Move to the next item

In deep mode, track each item's outcome so the deep-mode loop below and its cumulative report have data to aggregate. Assign one of the following statuses per item: pass (passed in isolation — full-suite verification may still revert it), abandoned (3 fix attempts exhausted), or bug-found (failure revealed a pre-existing bug in source code). Append any bugs found to accumulated-bugs. The per-item appends into accumulated-items happen in the Deep mode loop subsection below, after Final verification has settled the pass vs reverted — test failure status for each entry.

Final verification

After all tests are written, run the full test suite to ensure no regressions. Follow the verification protocol from $CLAUDE_PLUGIN_ROOT/skills/init/references/verification-protocol.md — run tests fresh, read complete output, and report actual results with evidence before claiming success.

In deep mode: if the full-suite run reveals that a newly-added test file causes regressions (either the new test itself failing under the full suite, or causing other tests to fail), revert that test file and promote its status from pass to reverted — test failure (to be appended by the Deep mode loop below). abandoned and bug-found statuses are not affected — only a pass is eligible for this promotion.

Deep mode loop

Normal mode and harness mode: Skip this subsection — normal mode proceeds to Step 5; harness mode proceeds to Step 6.

Interactive deep mode: After Final verification above completes for this iteration:

1. Append this iteration's results to accumulated-items — one entry per covered symbol (a test file exercising 3 functions yields 3 entries). Each entry uses the schema defined in Step 1 state init (file, target, iteration, status), with status settled during Per-test workflow and Final verification.
2. Update total-added by the count of entries with status pass.
3. Update total-reverted by the count of entries with status reverted — test failure.
4. Compute iteration-coverage-delta as the difference between current coverage and coverage at the start of this iteration (use 0 if the coverage tool is unavailable), then add it to accumulated-coverage-delta.

Then check termination conditions in order; the first matching condition stops the loop:

1. No new testable items discovered this iteration (convergence) → stop. Filter the discovery agent's Testability Classification Testable section against accumulated-items using exact-match on the file:function/class portion (format set in Step 1 state init). If every Testable entry is already a target in accumulated-items, convergence is reached. Report: "Deep mode complete — converged on iteration [N] with no remaining testable items."
2. No newly-written test passed verification — at least one test was written this iteration AND total-added for this iteration is 0 (every attempt ended in reverted — test failure, abandoned, or bug-found) → stop. Report: "Deep mode stopped — all tests added in iteration [N] caused failures."
3. Coverage plateau — coverage tool is available AND |iteration-coverage-delta| ≤ 0.5pp → stop. Report: "Deep mode stopped — coverage plateau on iteration [N]."
4. iteration-count >= the cap → cap reached. Report: "Deep mode reached the iteration cap ([cap]). Remaining testable items may exist — continue in a fresh conversation: re-run /optimus:unit-test deep, increase the cap with /optimus:unit-test deep [higher-cap], or narrow scope with /optimus:unit-test deep <scope>."
5. Otherwise → continue to the next pass.

In every outcome (stop or continue), present the iteration report immediately after the termination/continuation message. Informational and non-blocking — no user prompt follows:

#### Iteration [N] — Report

Cumulative coverage: [coverage-summary]

| # | File | Target | Status |
|---|------|--------|--------|
[one row per accumulated-items entry where iteration == current — one entry per covered symbol, per Step 1 state init]

Column definitions:

• # — Sequential number within this iteration
• File — Test file path (or — for reverted/abandoned)
• Target — the file:function/class being tested (format set in Step 1 state init)
• Status — the entry's state-enum value: pass, reverted — test failure, bug-found, or abandoned (as defined in Step 1 state init)

If the loop will continue (condition 5), after the iteration report also show the progress summary: "Iteration [N] of up to [cap] — [total-added] tests added so far, [total-reverted] reverted, cumulative coverage [coverage-summary]. Starting next pass..." If the next iteration will be 3 or higher, append: "Note: context is accumulating — if output quality degrades, consider finishing in a fresh conversation." Then increment iteration-count and return to Step 2 for the next discovery-and-write pass. Keep the same scope from Step 1.

After the loop ends (conditions 1–4 triggered), present a cumulative report in place of the Step 5 single-pass summary:

## Unit Test — Deep Mode Cumulative Report

**Summary:**
- Total iterations: [iteration-count]
- Total tests added: [total-added]
- Total tests reverted (failures): [total-reverted]
- Cumulative coverage delta: [coverage-summary]
- Final test status: pass / fail / not available

**All Tests Added:**

| # | Iter | File | Target | Status |
|---|------|------|--------|--------|
[one row per item from accumulated-items, across all iterations, ordered by iteration then sequence]

### Bugs Discovered
- [one bullet per entry in `accumulated-bugs`; omit this section if empty]

### Not Testable Without Refactoring
- [one bullet per entry in `accumulated-untestable`; omit this section if empty]
- To address these, run `/optimus:refactor testability` to prioritize testability improvements.

Then skip Step 5. Recommend running /optimus:refactor testability to prioritize testability improvements (or /optimus:refactor for balanced code quality review), or /optimus:tdd to continue development with test-driven workflow. Tell the user: Tip: for best results, start a fresh conversation for the next skill — each skill gathers its own context from scratch.

Step 5: Summary

Interactive deep mode: Skip this step — the cumulative report rendered by the deep mode loop at the end of Step 4 replaces the single-pass summary below.

Harness mode: Skip this step — Step 6 emits the structured JSON output instead.

Normal mode: Report to the user:

## Unit Test Summary

### Coverage
- Coverage tooling: [tool name / not configured]
- Before: [X]% → After: [Y]%
- Achievable target (without refactoring): ~[Z]%

### Tests Created
| # | File | Target | Status |
|---|------|--------|--------|
| 1 | src/__tests__/auth.test.ts | auth module exports | ✓ Pass |
| 2 | src/__tests__/validate.test.ts | validation utilities | ✓ Pass |
| ... | ... | ... | ... |

### Bugs Discovered
- [List of bugs found in existing code — reported, not fixed]

### Not Testable Without Refactoring
- [List of code flagged as untestable — with brief explanation of what structural change would be needed]
- To address these, run `/optimus:refactor testability` to prioritize testability improvements.

For multi-repo workspaces, present results per repo (one summary block per repo) and include the repo name/path in each section header.

Recommend running /optimus:refactor testability to prioritize testability improvements (or /optimus:refactor for balanced code quality review), or /optimus:tdd to continue development with test-driven workflow.

Tell the user: Tip: for best results, start a fresh conversation for the next skill — each skill gathers its own context from scratch.

Step 6: Harness Output (harness mode only)

If running under HARNESS_MODE_ACTIVE, output structured JSON instead of the Step 5 markdown summary. Output it in a json:harness-output fenced block and then stop — do not loop, do not present recommendations, do not use AskUserQuestion.

```json:harness-output
{
  "iteration": <cycle number from progress file>,
  "phase": "unit-test",
  "coverage": {
    "tool": "<coverage tool name or null>",
    "before": <percentage or null>,
    "after": <percentage or null>,
    "delta": <percentage or null>
  },
  "tests_written": [
    {
      "file": "<test file path>",
      "target_file": "<source file being tested>",
      "target_description": "<what it tests>",
      "test_count": <number of test cases>,
      "status": "<pass | fail-fixed | fail-abandoned>",
      "failure_reason": "<reason or null>"
    }
  ],
  "untestable_code": [
    {
      "file": "<source file path>",
      "line": <start line>,
      "end_line": <end line>,
      "function": "<function or class name>",
      "barrier": "<hardcoded-dependency | tight-coupling | global-state | ...>",
      "barrier_description": "<brief explanation>",
      "suggested_refactoring": "<what refactoring would help>"
    }
  ],
  "bugs_discovered": [
    {
      "file": "<path>",
      "line": <line>,
      "description": "<bug behavior>",
      "severity": "<High | Medium | Low>"
    }
  ],
  "no_new_tests": <true if zero new tests were written>,
  "no_untestable_code": <true if no untestable code was found>,
  "no_coverage_gained": <true if coverage delta is zero or negative>
}
```