Skim Master Rules — AI skill for Claude Code
Skim Master — output governance for AI agents.
How to install Skim Master Rules
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open chintheman/skim-master-rules and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Skim Master Rules does
Skim Master — output governance for AI agents. Nothing load-bearing is only in the middle: the first line alone is enough, first and last is enough, the rest adds detail without surprises. Claude Code · Claude app · Hermes
Alternatives in AI
- Firecrawl MCP Server — 🔥 Official Firecrawl MCP Server - Adds powerful web scraping and search to Cursor, Claude and any other LLM c 6.1k ★
- Surface — Show ranked attack surface for a target from its recon manifest + hunt memory 3.8k ★
- Axonhub — ⚡️ Open-source AI Gateway — Use any SDK to call 100+ LLMs 3.2k ★
README
Skim Master — output rules for AI agents
Two layers that stop AI assistants from writing walls of text:
- Always-on output governance — a rules block (Scope → Rule 0 → Rules → Pre-Send Check → Off-switches) that shapes every chat reply: answer first, plain English instead of system vocabulary, substance over format, no preamble, no recap, no closers.
- On-demand overlay — end any prompt with the word
skimand the agent reformats that one answer (kill the wall, number the steps, bold the load-bearing words).
Built from months of real corrections — every rule traces to a specific moment where a default AI answer wasted the reader's attention. Runs in production across Hermes Agent (core), Claude Code, and the Claude app.
What's inside
| Path | What it is |
|---|---|
rules/skim-master-rules.md |
Canonical rules block. Drop-in file for any agent with standing instructions (Claude Code, Hermes, custom system prompts) |
claude-app/instructions-for-claude.md |
Slim version for the Claude app's account-level "Instructions for Claude" (~1,500-char cap) |
examples/before-after-explained.md |
The same answer before/after, with the rule firing on every line explained |
SKILL.md |
The skim-suffix overlay as an installable skill (adhd still fires it) |
LICENSE |
MIT |
Install
**Claude Code** (governs every session):
mkdir -p ~/.claude/rules
cp rules/skim-master-rules.md ~/.claude/rules/
Files in `~/.claude/rules/` auto-load at session start — no import line, no CLAUDE.md edit. (You can also append the block to `~/.claude/CLAUDE.md` directly.) New sessions only; a running session keeps its start-of-session context.
**Claude app** (claude.ai):
Settings → General → Profile → "Instructions for Claude" → paste the slim version from `claude-app/instructions-for-claude.md`. Applies to new chats. For full governance on a work project, paste the complete block into that Project's instructions instead
Related Skills
Toggle
Toggle the statusline between minimal (identity + branch + model only) and detail (full multi-line) modes
Skillmgmnt
Read-only observability for AI-agent skills: catalog, load topology across 47+ tools, real usage, governance f
Large Codebase Command
AI Code Agent Taming 2026: Master 15K+ Line Repos Without Context Overflow Errors
Retornatus
Repo-native governance for AI coding agents (Cursor, Claude Code, Codex). Agree on the finish line in writing,
Bearing
An intel layer for AI coding agents that assumes its own rules might be wrong — and keeps the evidence. Author
Allocate
Stage 04. Invoke the solution-architect in allocate mode to place the intelligence — assign every operating-ma
Related Agents
Room Steward
Read the last N room messages and flag breaches of the room rules - overlong posts, method detail that belongs
Absorption Check
Warm per-round absorption check for design.md and track files: two-way coverage match between the research log
Act 3 Pivot
Read-only narrative steward for Act III (slide 15) — the pivot slide that turns the deck from workflows to age