Add a new rule, convention, or instruction to the project's agent configuration. Analyzes the rule and helps decide placement: root CLAUDE.md (universal rules), docs/agents/ files (topic-specific guidance), or a new skill (complex workflows). Use when users say: '/agent-add-rule', 'add a rule', 'add convention', 'new coding standard', 'add instruction for claude', 'update claude.md with'.
98
Quality
100%
Does it follow best practices?
Impact
98%
1.12xAverage score across 3 eval scenarios
Universal rule placement and token-cost discipline
Reads CLAUDE.md first
100%
75%
Rule in root CLAUDE.md
100%
100%
Not in docs/agents/
100%
100%
Placed under appropriate section
100%
100%
Rule is concise
100%
100%
Token cost mentioned
40%
100%
Universal classification reasoning
100%
100%
Checks existing skills list
100%
62%
No duplication introduced
100%
100%
Rule prohibits alternatives
100%
100%
recommendation.md present
100%
100%
Without context: $0.3359 · 1m 29s · 21 turns · 27 in / 4,122 out tokens
With context: $0.3323 · 1m 28s · 20 turns · 101 in / 4,050 out tokens
New docs/agents/ file with routing signal
New docs/agents/ file created
100%
100%
Rule NOT in root CLAUDE.md body
100%
100%
CLAUDE.md links section updated
100%
100%
Routing signal present
100%
100%
docs/agents/ file has clear heading
100%
100%
No cross-references between docs/agents/ files
100%
100%
requestId convention included
100%
100%
Error envelope convention included
100%
100%
Reads existing structure first
100%
100%
Semi-dynamic reasoning
100%
100%
Without context: $0.3267 · 1m 32s · 21 turns · 25 in / 5,001 out tokens
With context: $0.4641 · 1m 39s · 26 turns · 30 in / 5,207 out tokens
Workflow classified as skill, not rule file
Classified as skill
46%
100%
Recommends /agent-skill-creator
0%
100%
Workflow NOT added to CLAUDE.md
100%
100%
Workflow NOT added to docs/agents/
100%
100%
Skill content provided
80%
100%
Reads existing structure
100%
100%
Explains skill vs docs/agents/ distinction
40%
100%
Explains skill vs root distinction
100%
100%
Token cost acknowledged
100%
100%
Without context: $0.2272 · 1m 11s · 12 turns · 17 in / 3,404 out tokens
With context: $0.4457 · 2m 2s · 21 turns · 277 in / 5,798 out tokens
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.