Slop Cop plugin catches LLM prose tells in Claude Code and Cursor
Slop Cop is a Go CLI and plug-and-play plugin for Claude Code and Cursor that detects LLM-generated prose patterns across 48 rules, silently revising agent drafts before they're returned to the user.
Score breakdown
Integrate Slop Cop into your Claude Code or Cursor workflow to automatically self-edit AI-generated prose for 48 common LLM tell patterns — no manual review required.
- 01Slop Cop is a Go CLI tool authored by yasyfm that detects LLM-generated prose patterns and emits structured JSON reports for automated agent consumption.
- 02The tool evaluates prose against 48 detection rules via the `slop-cop check -` command.
- 03It ships as a plug-and-play plugin for both Claude Code and Cursor, silently revising agent drafts before they are returned to the user.
Yasyfm has published Slop Cop, a Go CLI tool and agentic plugin designed to catch the telltale rhetorical and structural patterns of LLM-generated prose. Unlike traditional linters aimed at human developers, Slop Cop outputs structured JSON reports intended for automated agent consumption — no TUI, no highlighting, no interactive prompts. It integrates directly into Claude Code and Cursor as a plug-and-play skill, intercepting the agent's own prose drafts and running them through `slop-cop check -` against 48 detection rules before the response is returned to the user.
Claude Code users can install via `/plugin marketplace add yasyf/slop-cop` followed by `/plugin install slop-cop@slop-cop`, while Cursor users install by pointing the Plugins panel at the Git URL.
Installation is frictionless: on first use, the skill runs `scripts/install-binary.sh` (or `.ps1` on Windows) to download a prebuilt platform binary from the latest GitHub Release, caching it inside the plugin directory for subsequent runs. Claude Code users can install via `/plugin marketplace add yasyf/slop-cop` followed by `/plugin install slop-cop@slop-cop`, while Cursor users install by pointing the Plugins panel at the Git URL. The skill activates automatically whenever the agent is asked to write or polish prose artifacts such as blog posts, docs, PR descriptions, commit messages, release notes, and emails. An end-to-end test script, `scripts/test-plugin.sh`, spawns a `claude -p` subshell with `--plugin-dir` pointed at a local checkout to verify activation.
Key facts
- 01Slop Cop is a Go CLI tool authored by yasyfm that detects LLM-generated prose patterns and emits structured JSON reports for automated agent consumption.
- 02The tool evaluates prose against 48 detection rules via the `slop-cop check -` command.
- 03It ships as a plug-and-play plugin for both Claude Code and Cursor, silently revising agent drafts before they are returned to the user.
- 04The binary auto-installs on first use via `scripts/install-binary.sh` (or `.ps1` on Windows), requiring no Go toolchain.
- 05The skill activates for prose tasks including blog posts, docs, PR descriptions, commit messages, release notes, and emails.
- 06Claude Code installation uses `/plugin marketplace add yasyf/slop-cop` and `/plugin install slop-cop@slop-cop`; Cursor installs via Git URL.
- 07An end-to-end test script `scripts/test-plugin.sh` verifies plugin activation using a `claude -p` subshell with `--plugin-dir`.