/endless-toil
Audible code-quality reactions for code reading, review, refactoring, debugging, and repository exploration. Use when the user wants Codex to scan code and play escalating groaning sounds based on suspicious code quality signals, especially requests like "make Codex suffer while
$ npx -y skills add andrewvos/endless-toil --skill endless-toil --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
- Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.Auto-invocation is when the right skill fires by itself at the right moment, driven by a FLOW.md router and a hook, instead of you invoking it by name. It is the difference between a skill being installed and a skill actually getting used.Read the full definition →
- You can call itInvoke it directly when you want it.
- Slash command
/endless-toil
Context preview
The summary Claude sees to decide when to auto-load this skill.
Audible code-quality reactions for code reading, review, refactoring, debugging, and repository exploration. Use when the user wants Codex to scan code and play escalating groaning sounds based on suspicious code quality signals, especially requests like "make Codex suffer while
SKILL.md
endless-toil.SKILL.mdname: endless-toil
description: Audible code-quality reactions for code reading, review, refactoring, debugging, and repository exploration. Use when the user wants Codex to scan code and play escalating groaning sounds based on suspicious code quality signals, especially requests like "make Codex suffer while reading this code", "play sounds based on code quality", "audible code review", or explicit use of $endless-toil.
Endless Toil
Overview
Use this skill to add audible reactions while inspecting code. After every time Codex reads code, immediately run the bundled scanner against the file, directory, or narrowest relevant path that was just inspected so the machine audibly groans through seven escalating levels of distress at what it encountered.
This skill cannot automatically hook every internal Codex file-read event outside a turn where the skill is active. When active, treat each code-read operation as requiring a follow-up sound reaction before continuing.
Quick Start
Resolve bundled files relative to the directory containing this `SKILL.md`; do not assume the current working directory is the skill directory. In examples below, `SKILL_DIR` means that directory.
Run:
python3 "$SKILL_DIR/scripts/endless_toil.py" .
Useful options:
python3 "$SKILL_DIR/scripts/endless_toil.py" src tests --max-sounds 8
python3 "$SKILL_DIR/scripts/endless_toil.py" . --dry-run
python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --threshold 3
python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --verbose
python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --foreground
Audition the sounds directly without scanning code:
python3 "$SKILL_DIR/scripts/test_sounds.py"
python3 "$SKILL_DIR/scripts/test_sounds.py" groan wail abyss
python3 "$SKILL_DIR/scripts/test_sounds.py" --list
Workflow
1. Identify the code paths relevant to the user's request. 2. Read code normally using the appropriate local tools. 3. After each code read, run `scripts/endless_toil.py` from this skill directory on the path just read, preferably with `--max-sounds 1` for single files. 4. Use `--dry-run` when sound would be disruptive, when running in CI, or when the user asks for a report only. 5. Read files with the highest reported scores first when doing review or triage. 6. Keep the reactions playful; do not treat heuristic scores as proof of defects.
After-Read Ritual
Whenever this skill is active and Codex reads code with commands like `sed`, `rg`, `cat`, `nl`, `git show`, or an editor/viewer tool, immediately play a reaction for the file or closest path that was read:
python3 "$SKILL_DIR/scripts/endless_toil.py" path/to/file --max-sounds 1
If the read covered multiple files, scan the smallest shared directory or the exact file list:
python3 "$SKILL_DIR/scripts/endless_toil.py" file_a.ts file_b.ts --max-sounds 3
Reaction Levels
- `silence`: no notable quality distress.
- `murmur`: the code has begun to trouble the room.
- `groan`: mild smell, such as TODO/FIXME density, debug logging, or long lines.
- `moan`: stronger concern, such as broad exception handling or many `any`/ignore directives.
- `wail`: accumulated structural pain, such as large files with high nesting.
- `howl`: risky execution paths, secrets, or a heavy cluster of suspicious signals.
- `shriek`: severe distress; multiple high-risk patterns are present.
- `abyss`: the deepest reaction, reserved for truly cursed score totals.
Requirements
- Python 3.10+
- A local audio player: `afplay`, `paplay`, `aplay`, or `ffplay`
If an audio player is unavailable, print the scan report and skip playback.
Real Sample Playback
Use real recorded samples only. The bundled `assets/sounds/zombie_moan_public_domain.ogg` sample is a public-domain recorded human-character moan from Wikimedia Commons, authored by Gregory Weir and sourced from PDSounds. The plugin ships seven pre-rendered `.wav` variants in `assets/sounds/generated/`.
Script Notes
The script uses simple static heuristics. For audio, it plays pre-rendered `.wav` files with the first available local player among `afplay`, `paplay`, `aplay`, or `ffplay`.
By default, scan commands queue audio playback in a detached background process so the agent thread can continue immediately. Use `--foreground` only when explicitly testing playback and waiting for the sound to finish is acceptable.
Normal scan commands are quiet by default so agent threads do not fill with reports. Use `--verbose` when you need scan details or playback status, and use `--dry-run` when you want a report without sound.
If no player is available, the script still prints the report and exits successfully unless scanning itself fails.
Installation Layout
Install this whole directory as `endless-toil`:
- Codex personal skill: `~/.codex/skills/endless-toil/SKILL.md`
- Claude personal skill: `~/.claude/skills/endless-toil/SKILL.md`
- Claude project skill: `.claude/skills/endless-toil/SKILL.md`
Keep `SKILL.md`, `scripts/`, `assets/`, and `agents/` together. The scripts derive asset paths from their own location, so the skill can be moved as a folder.
Read more
name: endless-toil description: Audible code-quality reactions for code reading, review, refactoring, debugging, and repository exploration. Use when the user wants Codex to scan code and play escalating groaning sounds based on suspicious code quality signals, especially requests like "make Codex suffer while reading this code", "play sounds based on code quality", "audible code review", or explicit use of $endless-toil.
Endless Toil
Overview
Use this skill to add audible reactions while inspecting code. After every time Codex reads code, immediately run the bundled scanner against the file, directory, or narrowest relevant path that was just inspected so the machine audibly groans through seven escalating levels of distress at what it encountered.
This skill cannot automatically hook every internal Codex file-read event outside a turn where the skill is active. When active, treat each code-read operation as requiring a follow-up sound reaction before continuing.
Quick Start
Resolve bundled files relative to the directory containing this `SKILL.md`; do not assume the current working directory is the skill directory. In examples below, `SKILL_DIR` means that directory.
Run:
python3 "$SKILL_DIR/scripts/endless_toil.py" .
Useful options:
python3 "$SKILL_DIR/scripts/endless_toil.py" src tests --max-sounds 8 python3 "$SKILL_DIR/scripts/endless_toil.py" . --dry-run python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --threshold 3 python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --verbose python3 "$SKILL_DIR/scripts/endless_toil.py" bad_file.ts --foreground
Audition the sounds directly without scanning code:
python3 "$SKILL_DIR/scripts/test_sounds.py" python3 "$SKILL_DIR/scripts/test_sounds.py" groan wail abyss python3 "$SKILL_DIR/scripts/test_sounds.py" --list
Workflow
1. Identify the code paths relevant to the user's request. 2. Read code normally using the appropriate local tools. 3. After each code read, run `scripts/endless_toil.py` from this skill directory on the path just read, preferably with `--max-sounds 1` for single files. 4. Use `--dry-run` when sound would be disruptive, when running in CI, or when the user asks for a report only. 5. Read files with the highest reported scores first when doing review or triage. 6. Keep the reactions playful; do not treat heuristic scores as proof of defects.
After-Read Ritual
Whenever this skill is active and Codex reads code with commands like `sed`, `rg`, `cat`, `nl`, `git show`, or an editor/viewer tool, immediately play a reaction for the file or closest path that was read:
python3 "$SKILL_DIR/scripts/endless_toil.py" path/to/file --max-sounds 1
If the read covered multiple files, scan the smallest shared directory or the exact file list:
python3 "$SKILL_DIR/scripts/endless_toil.py" file_a.ts file_b.ts --max-sounds 3
Reaction Levels
- `silence`: no notable quality distress.
- `murmur`: the code has begun to trouble the room.
- `groan`: mild smell, such as TODO/FIXME density, debug logging, or long lines.
- `moan`: stronger concern, such as broad exception handling or many `any`/ignore directives.
- `wail`: accumulated structural pain, such as large files with high nesting.
- `howl`: risky execution paths, secrets, or a heavy cluster of suspicious signals.
- `shriek`: severe distress; multiple high-risk patterns are present.
- `abyss`: the deepest reaction, reserved for truly cursed score totals.
Requirements
- Python 3.10+
- A local audio player: `afplay`, `paplay`, `aplay`, or `ffplay`
If an audio player is unavailable, print the scan report and skip playback.
Real Sample Playback
Use real recorded samples only. The bundled `assets/sounds/zombie_moan_public_domain.ogg` sample is a public-domain recorded human-character moan from Wikimedia Commons, authored by Gregory Weir and sourced from PDSounds. The plugin ships seven pre-rendered `.wav` variants in `assets/sounds/generated/`.
Script Notes
The script uses simple static heuristics. For audio, it plays pre-rendered `.wav` files with the first available local player among `afplay`, `paplay`, `aplay`, or `ffplay`.
By default, scan commands queue audio playback in a detached background process so the agent thread can continue immediately. Use `--foreground` only when explicitly testing playback and waiting for the sound to finish is acceptable.
Normal scan commands are quiet by default so agent threads do not fill with reports. Use `--verbose` when you need scan details or playback status, and use `--dry-run` when you want a report without sound.
If no player is available, the script still prints the report and exits successfully unless scanning itself fails.
Installation Layout
Install this whole directory as `endless-toil`:
- Codex personal skill: `~/.codex/skills/endless-toil/SKILL.md`
- Claude personal skill: `~/.claude/skills/endless-toil/SKILL.md`
- Claude project skill: `.claude/skills/endless-toil/SKILL.md`
Keep `SKILL.md`, `scripts/`, `assets/`, and `agents/` together. The scripts derive asset paths from their own location, so the skill can be moved as a folder.
Hear your agent suffer through your code. Endless Toil runs alongside your coding agent in real time, playing escalating recorded human groans as the code it reads starts to look more cursed.