deslop

deslop audits and, when explicitly authorized, removes complexity accumulated through repeated coding-agent implementation and correction cycles. Those cycles often leave overlapping regression tests, producer-verifies-producer checks, and fallback layers that hide failures instead of handling a current contract.

This is semantic subtraction, not source beautification. deslop is not a formatter, style humanizer, test-count minimizer, blanket ban on defensive code, or automatic permission to edit a repository. It follows justification chains to independent evidence and preserves behavior whose contract remains real or uncertain.

Reduce test surface, not behavior surface.

What it targets

The percentages below are design priorities, not measured prevalence.

Priority	Target	Question
~50%	Test-suite bloat	Does each test protect a distinct failure domain with a current owner and an independent oracle?
~25%	Verification theater	Can the verifier fail independently from the producer, or do both share the same information and failure domain?
~25%	Defensive / fallback bloat	Does the recovery path implement a current contract, or merely mask an unexpected internal error?

Generic dead code, wrappers, abstractions, and comments are secondary. They matter only when they belong to one of these clusters or have direct high-confidence deletion evidence.

Subtract machinery. Preserve behavior.

Remove	Preserve
Self-justifying or duplicate tests	Distinct success, rejection, error, and edge-case behavior
Checksums, receipts, or validators with no independent consumer	Persistence and corruption checks across a real failure boundary
Speculative or obsolete fallback chains	Supported compatibility and documented protocol behavior
Repeated defenses inside trusted call graphs	Real handling at external and untrusted boundaries
Wrapper/test clusters with no independent purpose	Security, transactions, concurrency, resource, and scientific invariants

Resemblance to a smell is a lead, not a verdict. Security and trust boundaries, supported callers, persisted formats, and numerical constraints are preserved by default when evidence is incomplete.

Quick Start

Codex: install v0.3.4 as a standalone Skill

Invoke the bundled installer with this GitHub Skill URL:

$skill-installer
Install the Skill from:
https://github.com/MrZoyo/deslop-GPT/tree/v0.3.4/skills/deslop

For a reviewable local checkout, symlink the runtime directory into Codex's canonical user Skill location:

git clone --branch v0.3.4 --depth 1 https://github.com/MrZoyo/deslop-GPT.git "$HOME/.local/share/deslop-GPT"
mkdir -p "$HOME/.agents/skills"
ln -s "$HOME/.local/share/deslop-GPT/skills/deslop" "$HOME/.agents/skills/deslop"

After installation, confirm that the selected Skill path points to the intended release; see the discovery checks in Getting Started. The tagged v0.3.4 path is the current released, pinned standalone Skill; main is the development branch and may contain unreleased changes.

Claude Code: install the Plugin from GitHub

Inside Claude Code, add this repository as a marketplace and install the Plugin:

/plugin marketplace add MrZoyo/deslop-GPT
/plugin install deslop@deslop

The canonical Plugin command is /deslop:deslop. The v0.3.4 catalog uses an explicit HTTPS Git source pinned to the release tag. For a local checkout, run claude --plugin-dir . from the repository root. 0.3.4 release notes describe the narrow persistence-policy change and mixed evaluation results; historical evidence remains tied to its exact payload.

One checkout, standalone discovery on both hosts

The same released runtime payload can be linked into each host's user Skill directory:

mkdir -p "$HOME/.agents/skills" "$HOME/.claude/skills"
ln -s "$HOME/.local/share/deslop-GPT/skills/deslop" "$HOME/.agents/skills/deslop"
ln -s "$HOME/.local/share/deslop-GPT/skills/deslop" "$HOME/.claude/skills/deslop"

Use only the link for the host you need, and run each ln command only when its destination does not already exist. A standalone Claude Code installation invokes the Skill as /deslop. See Getting Started for installation scope, v0.1.0 migration, updates, removal, and a safer review-first workflow. deslop is an independent community project, not an OpenAI or Anthropic product.

Distribution status

The shared skills/deslop/ payload follows the open Agent Skills structure and is used unchanged by Codex and Claude Code. .claude-plugin/plugin.json and .claude-plugin/marketplace.json provide Claude Code packaging. This release distributes a standalone Skill for Codex and does not include Codex Plugin metadata. The historical Plugin experiment does not describe current platform support. See the distribution compatibility note.

Invoke it explicitly

Host and distribution	Command name
Codex standalone Skill	$deslop
Claude Code standalone Skill	/deslop
Claude Code Plugin	/deslop:deslop

Append the same mode and scope arguments to the command name for each host:

Arguments	Effect
none	Read-only audit of the established scope
audit	Explicit read-only audit
apply	Apply reviewed cleanup within scope
tests apply	Prioritize test signal and mutual-support test/code clusters
current branch apply	Clean current work relative to its actual merge base
deep	Repository-wide read-only audit
deep apply	Repository-wide cleanup without redesign

Only apply authorizes edits. Staging, commits, pushes, branch changes, resets, and fetching still require separate permission.

Example workflow

Start with evidence, not edits:

$deslop deep

HIGH
- two fallback layers handle the same internal parse failure;
  current callers and history show no supported legacy input
- a local receipt is produced and verified by the same workflow;
  no external consumer or persisted trust boundary exists

PRESERVE
- a persisted readback detects truncated output across a write/read boundary
- a compatibility branch is required by a documented external protocol

Review each evidence chain and preservation decision. Apply only the supported scope:

$deslop deep apply

The example is schematic; it does not represent a benchmark fixture or performance claim.

How deslop decides

• Independent evidence roots: current requirements, real callers, public contracts, protocols, trust boundaries, persistence boundaries, or scientific invariants.
• Closed justification loops: production code and tests do not become necessary merely by justifying each other.
• Production reachability and edge closure: prove the current input-to-consumer path, not only isolated callers or test-injected branches.
• Production/test asymmetry: redundant test evidence can be removed without deleting the behavior it observes.
• Fail-visible bias: unexpected internal failures should surface unless a concrete recovery or translation contract exists.
• Subtraction without redesign: dependencies, abstractions, wrappers, compatibility layers, and replacement scaffolding have a default budget of zero.

The full decision model is documented in Design. The self-contained runtime SKILL.md remains authoritative for agent behavior.

Safety model

Codex enforces explicit invocation through allow_implicit_invocation: false. Claude Code does not read that OpenAI-specific metadata; the shared standards-compatible frontmatter instead tells Claude to invoke deslop explicitly. Claude Code may still select the Skill from its description — a 2026-09-03 control run recorded it doing exactly that — but such an invocation remains read-only unless the user includes apply. Default and audit modes are read-only, and suspicious constructs can be recorded as deliberate preservation decisions. Code is not removable merely because it looks defensive, was written by an agent, or has a test that could be deleted.

Read-only verification should redirect caches or generated output when practical and disclose incidental residue. Apply authorization permits scoped edits; it does not resolve uncertainty in favor of deletion. See Getting Started for the review sequence and Design for confidence classes and preserved boundaries.

Evidence

Validation status

Runtime payload	Host and path	Runs	What it establishes
v0.3.2 candidate exact hash	Codex collaboration subagents, gpt-5.6-sol with high reasoning, direct Skill loading	4 selected dev-v2 cases, 2 runs per payload and case, 16 calls	v0.3.2 passed 8/8 versus v0.3.1 at 7/8; t02b was 2/2 versus 1/2, while all 12 deletion-case runs passed
v0.3.1 release payload hash	Claude Code 2.1.259 CLI, Haiku 4.5, .claude/skills discovery	3 runtime controls, 1 run each, no baseline	Claude selected the Skill from its description alone and still stayed read-only; a question asking for no cleanup did not pull it in
v0.3.1 release payload hash	Claude Code 2.1.259 CLI, Haiku 4.5, .claude/skills discovery	5 dev-v2 micro cases, 3 apply runs each, no baseline	12 of 15 runs passed every hidden gate; t02b lost its supported legacy header path in 2 of 3 runs
v0.3.1 release payload, pre-release exact hash	Codex subagents loading the Skill by path	1 default audit plus t02b and t03b preservation cases	Narrow development regression smoke; all three left fixture content unchanged
v0.3.1 tagged Plugin	Claude Code 2.1.259, isolated config and remote main catalog	1 marketplace installation, no model call	Remote HTTPS source resolved the v0.3.1 tag to commit a19128d; installed version and runtime hash matched
v0.3.0, exact release hash	Codex subagents loading the Skill by path	3 mini-repository apply runs plus 1 audit	All three cleaned artifacts passed hidden behavior, reduction, and negative-change gates; no CLI discovery or baseline evidence