github.com/peteromallet/desloppify (link)

---
name: desloppify
description: >
  Multi-language codebase health scanner. Use when the user explicitly asks
  to run desloppify, scan for technical debt, get a health score, or create
  a cleanup plan. Do NOT trigger for general code review, renaming, or
  fixing individual bugs.
---

<!-- desloppify-begin -->
<!-- desloppify-skill-version: 7 -->

# Desloppify

## 1. Your Job

Maximise the **strict score** honestly. Your main cycle: **scan → plan → execute → rescan**. Follow the scan output's **IN

---
name: desloppify
description: >
  Multi-language codebase health scanner. Use when the user explicitly asks
  to run desloppify, scan for technical debt, get a health score, or create
  a cleanup plan. Do NOT trigger for general code review, renaming, or
  fixing individual bugs.
---

<!-- desloppify-begin -->
<!-- desloppify-skill-version: 7 -->

# Desloppify

## 1. Your Job

Maximise the **strict score** honestly. Your main cycle: **scan → plan → execute → rescan**. Follow the scan output's **IN

## Hermes Agent Overlay

Hermes has built-in parallel subagent support via `delegate_task` (up to 3
concurrent children). Use `delegate_task(tasks=[...])` for subjective review
batches and per-stage triage support; avoid the older worktree-based guidance
here.

### Review workflow

1. Prepare review prompts and the blind packet:
   ```bash
   desloppify review --run-batches --dry-run
   ```
   This generates one prompt file per batch in
   `.desloppify/subagents/runs/<run-id>/prompts/` and prints the run di

## Hermes Agent Overlay

Hermes has built-in parallel subagent support via `delegate_task` (up to 3
concurrent children). Use `delegate_task(tasks=[...])` for subjective review
batches and per-stage triage support; avoid the older worktree-based guidance
here.

### Review workflow

1. Prepare review prompts and the blind packet:
   ```bash
   desloppify review --run-batches --dry-run
   ```
   This generates one prompt file per batch in
   `.desloppify/subagents/runs/<run-id>/prompts/` and prints the run di

## Claude Code Overlay

Use Claude subagents for subjective scoring work. **Do not use `--runner codex`** — use Claude subagents exclusively.

### Review workflow

Run `desloppify review --prepare` first to generate review data, then use Claude subagents:

1. **Prepare**: `desloppify review --prepare` — writes `query.json` and `.desloppify/review_packet_blind.json`.
2. **Launch subagents**: Split the review across N parallel Claude subagents (one message, multiple Task calls). Each agent reviews a subse

## Claude Code Overlay

Use Claude subagents for subjective scoring work. **Do not use `--runner codex`** — use Claude subagents exclusively.

### Review workflow

Run `desloppify review --prepare` first to generate review data, then use Claude subagents:

1. **Prepare**: `desloppify review --prepare` — writes `query.json` and `.desloppify/review_packet_blind.json`.
2. **Launch subagents**: Split the review across N parallel Claude subagents (one message, multiple Task calls). Each agent reviews a subse

## Gemini CLI Overlay

Gemini CLI has experimental subagent support, but subagents currently run
sequentially (not in parallel). Review dimensions one at a time.

### Setup

Enable subagents in Gemini CLI settings:
```json
{
  "experimental": {
    "enableAgents": true
  }
}
```

Optionally define a reviewer agent in `.gemini/agents/desloppify-reviewer.md`:

```yaml
---
name: desloppify-reviewer
description: Scores subjective codebase quality dimensions for desloppify
kind: local
tools:
  - read_file
  - se

## Gemini CLI Overlay

Gemini CLI has experimental subagent support, but subagents currently run
sequentially (not in parallel). Review dimensions one at a time.

### Setup

Enable subagents in Gemini CLI settings:
```json
{
  "experimental": {
    "enableAgents": true
  }
}
```

Optionally define a reviewer agent in `.gemini/agents/desloppify-reviewer.md`:

```yaml
---
name: desloppify-reviewer
description: Scores subjective codebase quality dimensions for desloppify
kind: local
tools:
  - read_file
  - se

## Qwen Code Overlay

Qwen Code loads skills from `.qwen/skills/<name>/SKILL.md` in a project or `~/.qwen/skills/<name>/SKILL.md` globally.

Install the project skill with:

```bash
desloppify update-skill qwen
```

Install the global skill with:

```bash
desloppify setup --interface qwen
```

Use the standard review workflow from the base skill. Automated `--runner qwen` batch reviews are not implemented yet; use the prepared packet/manual import workflow or another supported batch runner.

<!-- desloppify

## Qwen Code Overlay

Qwen Code loads skills from `.qwen/skills/<name>/SKILL.md` in a project or `~/.qwen/skills/<name>/SKILL.md` globally.

Install the project skill with:

```bash
desloppify update-skill qwen
```

Install the global skill with:

```bash
desloppify setup --interface qwen
```

Use the standard review workflow from the base skill. Automated `--runner qwen` batch reviews are not implemented yet; use the prepared packet/manual import workflow or another supported batch runner.

<!-- desloppify

## Codex Overlay

This is the canonical Codex overlay used by the README install command.

1. Prefer first-class batch runs: `desloppify review --run-batches --runner codex --parallel --scan-after-import`.
2. The command writes immutable packet snapshots under `.desloppify/review_packets/holistic_packet_*.json`; use those for reproducible retries.
3. Keep reviewer input scoped to the immutable packet and the source files named in each batch.
4. If a batch fails, retry only that slice with `desloppify review

## Codex Overlay

This is the canonical Codex overlay used by the README install command.

1. Prefer first-class batch runs: `desloppify review --run-batches --runner codex --parallel --scan-after-import`.
2. The command writes immutable packet snapshots under `.desloppify/review_packets/holistic_packet_*.json`; use those for reproducible retries.
3. Keep reviewer input scoped to the immutable packet and the source files named in each batch.
4. If a batch fails, retry only that slice with `desloppify review