Releases·Tool Release
Developer releases open-source 'claudish' skill to curb AI writing tropes in technical docs
A developer released a custom skill, claudish, on GitHub to curb recurring AI writing tropes in technical documentation. It targets habits like overly literary prose and contrastive framing, with goals similar to tropes.fyi. The author built it iteratively by adding corrections to a do-not-do list and invites pull requests.
4 engagement1 sourceMon, Sep 14, 2026, 07:06 PM
AI Overview
- cmdcolin's GitHub skill 'claudish' targets AI tropes like overly literary prose and 'contrastive framing', built from a 'DO_NOT_DO' list, with PRs welcome
Entities
Related
Social·Community
Any good Agent SKILLS.md for clear technical writing with a good tone
A developer asks for Agent SKILLS.md files that make agents write with a clear, human technical tone. Self-made attempts, including having an LLM analyze good documentation samples, failed to improve writing quality enough.
A developer asks for Agent SKILLS.md files that make agents write with a clear, human technical tone. Self-made attempts, including having an LLM analyze good documentation samples, failed to improve writing quality enough.
31 engagement1 sourceredditWed, Sep 30, 2026, 08:59 AM
Releases·Tool Release
AI Slop Detector: GitHub repo catalogs 120 AI-writing patterns with skills for Claude Code and Codex
FeedHive published an open-source AI Slop Detector cataloging 120 writing patterns that make text read as AI-generated, organized into 9 categories with explanations, bad examples, and rewrites. It ships as an agent skill for Claude Code and Codex, a browser-based checker, and structured JSON for custom tooling. The tool targets style cleanup rather than authorship detection, which matters for practitioners trying to keep AI-assisted output readable.
FeedHive published an open-source AI Slop Detector cataloging 120 writing patterns that make text read as AI-generated, organized into 9 categories with explanations, bad examples, and rewrites. It ships as an agent skill for Claude Code and Codex, a browser-based checker, and structured JSON for custom tooling. The tool targets style cleanup rather than authorship detection, which matters for practitioners trying to keep AI-assisted output readable.
39 engagement1 sourcegithubMon, Oct 12, 2026, 09:38 AM
Products·launched
Writ: open-source LLM agent skill that self-edits text using Wikipedia's 'Signs of AI Writing' checklist
Writ is an open-source agent skill that turns Wikipedia's community guide 'Signs of AI Writing' into a self-edit checklist for LLMs. Agents use it to detect and fix telltale AI writing patterns before publishing. It targets writers and developers whose LLM-generated text gets flagged or rejected.
Writ is an open-source agent skill that turns Wikipedia's community guide 'Signs of AI Writing' into a self-edit checklist for LLMs. Agents use it to detect and fix telltale AI writing patterns before publishing. It targets writers and developers whose LLM-generated text gets flagged or rejected.
8 engagement1 sourceredditSat, Sep 5, 2026, 11:16 AM
Products·launched
yomiyasu: Agent Skill that rewrites AI-generated Japanese into natural, human-sounding prose
yomiyasu is an Agent Skill for AI coding environments that revises AI-generated Japanese text, removing unnatural patterns and rewriting it into natural, information-dense Japanese. The LLM acts as the editing engine, guided by an analysis of linguistic artifacts typical of AI writing. It targets developers and writers using tools like Codex, Claude Code, and Cursor who want to de-flavor AI Japanese output.
yomiyasu is an Agent Skill for AI coding environments that revises AI-generated Japanese text, removing unnatural patterns and rewriting it into natural, information-dense Japanese. The LLM acts as the editing engine, guided by an analysis of linguistic artifacts typical of AI writing. It targets developers and writers using tools like Codex, Claude Code, and Cursor who want to de-flavor AI Japanese output.
197 engagement1 sourcegithubWed, Sep 30, 2026, 11:49 AM
Products·launched
Claude skills: reusable rule sets with triggers and refusals
A system for turning ad-hoc Claude instructions into reusable 'skills' that define trigger conditions, refusal behaviors, and optional JSON output. The author shares a free anti-slop skill that bans common AI writing tells. It solves the problem of re-explaining specific rules every session for Claude power users.
A system for turning ad-hoc Claude instructions into reusable 'skills' that define trigger conditions, refusal behaviors, and optional JSON output. The author shares a free anti-slop skill that bans common AI writing tells. It solves the problem of re-explaining specific rules every session for Claude power users.
17 engagement1 sourceredditWed, Jul 15, 2026, 09:10 PM