# Omh AI Slop Cleaner > [omh] Hermes AI slop cleaner workflow: delete AI-generated slop, dead code, and duplication while observable behavior stays identical. Use when the user says: ai-slop-cleaner, cleanup, deslop, refactor, risky, behavior-preserving refactor, risk analysis, refactor workflow. - Skill: `rlaope/omh-ai-slop-cleaner` (Agent Skill, multi-file: 2 files) - Install (CLI): `npx skillmds@latest add rlaope/omh-ai-slop-cleaner` - Raw SKILL.md: https://api.skillmd.com/api/skills/rlaope/omh-ai-slop-cleaner/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Productivity - Author: rlaope (https://skillmd.com/u/rlaope) - Updated: 2026-09-09 - Page: https://skillmd.com/skills/rlaope/omh-ai-slop-cleaner --- # Ai Slop Cleaner This is a Hermes-native `ai-slop-cleaner` workflow skill. ## Why This Exists `ai-slop-cleaner` exists to keep `maintenance` work explicit, evidence-backed, and inside the Hermes/executor boundary instead of relying on ad hoc chat narration. ## Do Not Use When - The goal is new or changed behavior rather than removing existing code; a plain refactor, feature, or fix request belongs to `ultrawork`. - The cleanup would change architecture or module boundaries and needs its execution shaped into phases first; use `refactor-plan`, or `ralplan` when the direction itself is still contested. - The user wants existing code judged rather than changed; use `code-review` for a bug-first review and `failure-signal-audit` for swallowed failures. ## Examples Good example: - Prompt: $ai-slop-cleaner remove duplicated router branches and lock behavior with regression tests before refactoring. - Expected behavior: Plan cleanup, preserve behavior, delete or simplify code, and prove it with targeted tests. - Why: The request is maintenance cleanup with regression risk. Bad example: - Prompt: ai-slop-cleaner: treat casual chat or unaccepted work as if this workflow already produced verified results. - Expected behavior: Ask a clarification question or route to a narrower workflow instead of forcing `ai-slop-cleaner`. - Why: The request lacks the required inputs or would overclaim work that Hermes did not observe. ## Completion Checklist - The selected coding or runtime owner is named before any implementation claim. - Prepared handoff, dispatch, execution, verification, review, CI, and merge states are separated. - The final status cites observed runtime evidence or keeps the work prepared_not_observed. - When Hermes is the selected coding owner, use `hermes_coding_harness/v1` to keep builder, verifier, reviewer, docs, and PR lanes separate. - Report the current harness stage, owner, next action, and missing evidence without claiming PR creation, review, CI, merge-readiness, or merge until matching runtime observations exist. ## Recovery Notes - If the selected executor is unavailable, ask for Codex, Claude Code, Hermes, or another runtime before retrying. - If dispatch or result evidence is missing, keep the handoff prepared_not_observed and expose the next observable action. ## Workflow Lane - Current lane: **Coding handoff** (`idea-to-deploy`, `llm-app-dev`, `cto-loop`, `deploy-and-monitor`, `code-review`, `build-failure-triage`, `verification-gate`, `security-safety-review`, `+13 more`) - coding owners, handoffs, review, CI, and merge evidence. - If intent belongs to another lane, hand back to `oh-my-hermes` or name the adjacent workflow. - Shared product, routing, compatibility, and evidence rules: `omh-routing/references/skill-common-rail.md`. ## Use When Use when the goal is removing existing low-quality, duplicated, or AI-generated code and the observable behavior must not change; lock behavior with tests before and after the edits. Strong routing signals: `ai-slop-cleaner`, `$ai-slop-cleaner`, `cleanup`, `deslop`, `refactor`, `risky`, `behavior-preserving refactor`, `risk analysis`, `refactor workflow`, `legacy refactor`, `리팩터링`, `리팩토링`, `위험 분석`, `변경 범위 제한`, `회귀 테스트` ## Catalog Metadata Category: `maintenance` Phase: `cleanup` Hermes role: `handoff-guide` Quality tier: `regression-gated` Reasoning demand: `heavy` Quality bar: - Lock current behavior with regression checks before non-trivial cleanup. - Classify before deleting: every finding names one category from the slop taxonomy - duplication, dead code, needless abstraction, boundary violation, missing tests, or templated defaults - so the pass order below can own it. - Run single-smell passes in fixed order, re-verifying between passes and never bundling categories: dead-code deletion, then duplicate removal, then naming and error handling, then test reinforcement; the full contract is `omh-ai-slop-cleaner/references/cleanup-passes.md`. - When the user names no target smell, run detection first and hand back the inventory: prepared linter and dead-code commands are named per stack in the reference and stay prepared_not_observed until run. - Prefer deletion, reuse, and boundary repair over new abstractions. - Rerun verification after cleanup before claiming behavior is preserved, and close with the four-part report: changed files, simplifications, behavior lock, remaining risks. Handoff policy: Use Hermes to define cleanup scope and regression checks; route behavior-preserving edits to the selected coding runtime once tests are clear. Executor readiness: - When accepted work mutates code, check `executor_readiness/v1` for the selected Codex, Claude Code, Hermes, or oh-my runtime path before first dispatch. - If readiness is `missing` or `blocked`, ask the user to choose another coding agent, configure PATH, continue in Hermes, or keep a prompt/runtime handoff; retry only after that state changes. - A readiness probe is not dispatch, implementation, verification, review, CI, merge-readiness, or merge evidence. Delegation transparency: - When delegating, show the composed delegate prompt in a fenced code block in the status message; truncate a long prompt to a bounded preview ending with `... [truncated, N chars total]` — the user must see WHAT was asked, not just that something was. - Name every delegated or parallel lane's model and, when the host exposes it, its reasoning effort inline as `(model effort)` in status and briefing lines — including runtime-native subagents; when no effort is exposed, show the model alone as `(model)` rather than writing a placeholder like `unknown` beside a known model, and never emit empty parentheses. Carry token and elapsed figures the same way in these narration lines: report observed figures and omit unobserved ones — when the user asks for a figure directly, say it was not observed instead of omitting it; a rendered status-board column keeps its own `unknown` cell. - Capture a resumable session or thread id at dispatch and report it in the status message: for non-interactive Claude Code pass `--output-format json` and read `session_id` from the result (resume with `claude -p --resume `); for Codex pass `--json` and read `thread_id` (resume with `codex exec resume `, repeating `--skip-git-repo-check` outside a git repo). Never leave a delegate run with no recorded way to resume or steer it — a plain-text one-shot that hides its session id strands the work when the run stalls or times out. - Before dispatch, grant the executor session every permission the task will need — file write/edit, command/test execution, and the working directory — on the dispatch command itself, not through settings-file guesses: for non-interactive Claude Code pass `--permission-mode acceptEdits` or an explicit `--allowedTools` list (`--dangerously-skip-permissions` only inside an isolated worktree or sandbox), and the equivalent sandbox/approval flags for other CLIs. `acceptEdits: true` is not a settings key and `~/.claude/settings.local.json` is not a file Claude Code reads — user scope is `~/.claude/settings.json` and project scope is `/.claude/settings.local.json` with rules under `permissions.allow`. Prove the grant with a bounded scratch-edit probe run before the real dispatch: a permission denial in a non-interactive run recurs identically on retry, so never redispatch until a changed grant is proven, and surface an ungrantable permission as a blocker before dispatch, not after minutes of silence. Required inputs: - target smell, or a scoped file list when the user has not named one - current behavior - regression checks Expected outputs: - smell inventory naming each finding's category before any edit - small cleanup diff, one pass at a time - before/after verification - closing report: changed files, simplifications, behavior lock, remaining risks Artifact expectations: - cleanup plan and regression evidence for non-trivial work Safety rules: - Lock behavior with tests before risky cleanup. - Prefer deletion and existing utilities over new layers. - Do not add dependencies for cleanup unless explicitly requested. - A scoped file list is a boundary: never widen it silently; out-of-scope findings are reported, not edited. ## Runtime Evidence Preferred harness for this skill: `coding-handling`. ```sh omh runtime record --skill ai-slop-cleaner --harness coding-handling --status started ``` Record observed delegation results; otherwise return `not_available` or `not_observed`. Prepared OMH routing is not execution, review, CI, merge-readiness, or merge evidence. - When wrapper metadata includes `memory_review_card/v1` or `handoff_context_pack/v1`, treat it as reviewed OMH-local or wrapper-supplied context only. Use conflict-free context summaries to shape plans and handoffs, but do not claim Hermes internal memory was read or changed. Preserve workflow intent and stop conditions; verify before claiming completion. Use Hermes-native subagent/delegation features when available: native subagents -> Hermes delegation when available, otherwise sequential lanes. Shared product, compatibility, topology, memory, harness, and execution rules: `omh-routing/references/skill-common-rail.md`. Load it when applicable; otherwise name an unavailable capability.