DocsCliAuto

quell auto

Non-interactive version of quell fix. Runs the full generate → verify → write loop for every survived mutant without asking for confirmation.

Usage

quell auto [OPTIONS]

Options

OptionDefaultDescription
--tool, -tmutmutSource tool: mutmut or stryker
--root, -r.Project root directory
--llmNoneLLM provider: anthropic, openai, ollama
--dry-runFalsePreview without writing any files

Examples

# Auto-fix all mutants
quell auto

# Preview what would be written
quell auto --dry-run

# With LLM for unknown operators
quell auto --llm anthropic

# From a monorepo sub-package
quell auto --root packages/payments

Output

Auto-fixing 14 survived mutants ...

  [1/14]  BOUNDARY_SHIFT   calculator.py:20   ✓ killed  → written
  [2/14]  ARITHMETIC_SWAP  payments.py:41     ✓ killed  → written
  [3/14]  RETURN_MUTATION  payments.py:55     ✓ killed  → written
  [4/14]  LOGICAL_SWAP     auth.py:88         ✓ killed  → written
  [5/14]  UNKNOWN          auth.py:104        ✗ skipped  (no LLM configured)
  ...
  [14/14] BOUNDARY_SHIFT   utils.py:201       ✓ killed  → written

  ─────────────────────────────────────────────────────────
  12 tests written   1 skipped (UNKNOWN)   1 skipped (verification failed)
  Score: 71% → 93%  (+22%)

Dry run

With --dry-run, Quell performs all generation and verification steps but writes nothing to disk. Useful for auditing what Quell would do before committing.

quell auto --dry-run

# Output shows "→ would write" instead of "→ written"

quell auto writes directly without confirmation. Use --dry-run first on a new project to understand what will be written.