Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsBlogPro
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Authors
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges
  • Chrome Extension
  • Skill Manager

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

ProTermsPrivacyRefunds
Back to skills

Dim Rules

ASecurity

Maintain the standing instructions agents load, machine-wide or in one repo, and the tool files that import them. Use when adding, sharpening or removing a rule, not when writing a doc that argues for one.

5 stars
0 votes
0 copies
0 views
Added 10/5/2026
ai-agentsgogit

Security Analysis

A100/100

Scanned 10/5/2026

$npx -y skills add cniska/dim-factory --skill dim-rules --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Dim Rules?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Dim Rules
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/cniska-dim-rules/badge)](https://www.skillsdirectory.com/skills/cniska-dim-rules)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
Files
SKILL.md
---
name: dim-rules
description: Maintain the standing instructions agents load, machine-wide or in one repo, and the tool files that import them. Use when adding, sharpening or removing a rule, not when writing a doc that argues for one.
argument-hint: "<the rule you want to land>"
---

# Rules

A rule costs tokens on every session that loads it and changes nothing on the sessions that do not. So the question is never whether the rule is true. It is which layer it belongs in, whether a gate could hold it instead of a sentence, and whether it is already written somewhere this reader will reach.

## Before writing a line

1. Is this rule already settled? `dim query search "<words the rule would use>"` finds where it was said, across every session. Find what settled it and sharpen that instead.
2. Is it being broken? The same search finds each time the owner had to say it again. A rule nobody breaks is a line paid for on every session to prevent nothing. A rule broken repeatedly under one skill belongs in that skill, not in the file every session loads.
3. What does the repo already do? `git log` reads the commit format off the repo's own history. What a repo does is the rule; a file that states something else is the thing that is wrong.
4. Does the concept already have a word? Read the project's glossary, where it keeps one. A rule that introduces a second word for a thing already named costs more than it states, because from then on both words are searched and only one is found.

## Choose the layer

A rule lives in one of these, and the cost falls as you go down:

- The machine-wide rules, read on every session in every repo. The most expensive line there is, so it holds only what is true of all work everywhere.
- One repo's rules, read on every session in that repo. Only what holds there and nowhere else; a general convention restated here is paid for twice and drifts.
- A station skill, read when that station runs. A rule governing one operation arrives exactly when it applies.
- A doc, read when the argument is needed. The reasoning, the measurement, the case that was weighed, never the rule itself.
- The record, read by a query. A fact, not a rule: what was decided, what a number was, which version shipped.

Put the rule in the lowest layer that still reaches the work. A rule that holds everywhere and sits in one repo is a rule the other repos break for free; a rule that holds in one repo and sits machine-wide is a line every unrelated session pays for.

## Prefer a gate to a sentence

Before writing prose, ask what could refuse. A subject's length, a schema version, an exit code, a shape rather than a meaning: each is a gate, and a gate holds whether or not anything was read. A sentence holds only when a model read it and chose to comply. Where the rule needs judgement, it stays a sentence. Where it does not, the sentence names the gate still to build.

## Merge, never append

Read the whole file first. A rule overlapping an existing one rewrites that one. A rule that generalizes several replaces them. A rule stated in two places is deduped to the one place its reader will reach. A file that only ever grew is a file nobody finished reading.

## Done when

- the rule sits in the lowest layer that reaches the work
- nothing was appended that an existing line could have absorbed
- every word in it is the glossary's, or the glossary now holds the new one
- what is mechanical is a gate, or is named as a gate still to build
- the file is shorter than it would have been, or the growth bought something no existing line carried

Attribution

cniskacniska
View sourceSee grades on GitHubMore from cniska →
SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Ship a skill? Prove it's safe.

Free 120-pattern security scan, letter grade, and an embeddable README badge.

Submit a skill

Related Skills

Caveman

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens". Stays on until "stop caveman" or "normal mode".

1100021 votes

Hyperplan

Adversarial multi-agent planning skill. Self-orchestrates 5 hostile category members (unspecified-low, unspecified-high, deep, ultrabrain, artistry) via team-mode for ruthless cross-critique debate, distills only the defensible insights, then MANDATORILY hands the distilled insight bundle to the `plan` agent for executable plan formalization. Use when planning needs maximum rigor and surfacing of weak assumptions, blind spots, and over-engineering. Triggers: 'hyperplan', 'hpp', '/hyperplan', ...

698431 votes

Writing Skills

Create and manage Claude Code skills in HASH repository following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns), UserPromptSubmit hook, and the 500-line rule. Includes validation and debugging with SKILL_DEBUG. Examples include rust-error-stack, cargo-dep...

3931 votes

Mcp Code Execution

Routes multi-tool workflows through MCP servers for large datasets and pipelines. Use when Bash tool overhead is limiting throughput on data-heavy tasks.

3421 votes

catchup

Recovers the conversation and failed tool calls of a previous Codex, Amp, Claude Code, Antigravity, Cline, Copilot CLI, Cursor, DeepSeek Harness, Grok Build, Kimi, OpenCode, Pi Agent, or ZCode session. Use when the user says "catch up", "what did the last session do", "get me up to speed", "I switched agents", asks to recover/summarize a previous session before continuing, or asks to diagnose or report a catchup failure. Do NOT use for the current conversation, git history, or any non-agent log.

741 votes
View all in ai-agents →