# Anti Slop Fix

> Runs the anti-slop audit on source code files and automatically applies fixes for detected issues. Invokes the anti-slop analysis first, then fixes each issue in-place. Use when the user wants to clean up AI slop automatically, fix slop patterns, or asks "fix slop", "auto-fix slop", "anti-slop fix". NOT when the user only wants the report (anti-slop) and not for whole-repo remediation from an audit's findings — that is a fix-plan execution, not a slop sweep.

- Skill: `b4r7x/anti-slop-fix` (Agent Skill)
- Install (CLI): `npx skillmds@latest add b4r7x/anti-slop-fix`
- Raw SKILL.md: https://api.skillmd.com/api/skills/b4r7x/anti-slop-fix/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: b4r7x (https://skillmd.com/u/b4r7x)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/b4r7x/anti-slop-fix

---


# Anti-Slop Fix

Runs the anti-slop audit and automatically applies fixes to detected issues.

## Workflow

### Step 1 — Run the audit

Execute the full anti-slop analysis workflow (Steps 1-5 from the `anti-slop` skill) on the provided file paths or globs. Read `../anti-slop/references/slop-code-patterns.md` for the pattern catalog. If that path does not exist (single-skill install), invoke the `anti-slop` skill directly instead; if it is not installed at all, stop and tell the user this skill depends on it.

Store the results internally — do NOT output the full report yet.

### Step 2 — Classify fixability

For each issue found, classify it:

- **Auto-fixable** — the fix is mechanical and safe:
  - Deleting unnecessary comments (Cat 1a, 1b, 1c, 1d, 1e)
  - Removing redundant else-after-return (Cat 3c)
  - Removing `=== true` / `=== false` comparisons (Cat 7b)
  - Removing unnecessary `async`/`await` on return (Cat 7c)
  - Removing unused imports (Cat 5a)
  - Removing unreachable code (Cat 5c)
  - Replacing ternary with `??` (Cat 7a)
  - Removing `return condition ? true : false` → `return condition` (Cat 7b)

- **Semi-auto** — likely safe but needs a sanity check:
  - Inlining single-use helper functions (Cat 2a)
  - Removing fallback values on non-nullable types (Cat 3e)
  - Shortening verbose error messages (Cat 3d)
  - Removing type assertions on already-typed values (Cat 6c)
  - Renaming "enhanced/optimized" variables (Cat 4a)

- **Manual only** — too risky to auto-fix, report these instead:
  - Fixing `as any` casts — requires understanding the correct type (Cat 6a)
  - Removing `@ts-ignore` — need to verify the fix compiles (Cat 6b)
  - Deleting exported-but-unused symbols — other repos may depend on them (Cat 5d)
  - Cross-file refactors (inlining across files)

### Step 3 — Apply auto-fixes

Process fixes **bottom-up within each file** (highest line number first) to preserve line numbers for subsequent edits.

For each auto-fixable issue:
1. Use the Edit tool to apply the fix
2. Record what was changed

Do NOT apply semi-auto fixes without telling the user. List them separately.

### Step 4 — Apply semi-auto fixes (with disclosure)

For each semi-auto issue:
1. Briefly state what will change and why
2. Apply the fix
3. If the fix is wrong or unclear, skip it and move it to "manual review"

### Step 5 — Output summary

```
## Anti-Slop Fix Report

**Files processed:** N
**Issues found:** N
**Auto-fixed:** N
**Semi-auto fixed:** N
**Manual review needed:** N

---

### Changes Applied

#### `path/to/file.ts`
- L12: Deleted comment restating code (`// set the name`)
- L45-52: Inlined `formatDate()` — was called once at L78
- L89: Replaced `isActive === true` with `isActive`

#### `path/to/other-file.ts`
- L3: Removed unused imports: `useCallback`, `useMemo`
- L67: Removed redundant `else` after `return`

---

### Manual Review Needed

| File | Line | Issue | Why manual |
|------|------|-------|------------|
| `api.ts` | 34 | `as any` cast | Need to determine correct response type |
| `types.ts` | 12 | `UserDTO` exported but unused | May be consumed by external packages |

---

### Verify

<lint + test output>
```

After applying fixes, run the project's narrowest matching lint and test commands yourself (from package.json scripts or the repo's docs) and include their output in the report. Only when no such commands exist, tell the user what to run manually. Never end the report with unverified fixes presented as done.

If there were no issues to fix:
```
## Anti-Slop Fix Report

**Files processed:** N
**Issues found:** 0

Clean. Nothing to fix.
```