rule-placement

Determines the correct context-delivery channel for generalized agent behavior rules.

3|1|Updated Dec 9, 2025
One-click install
npx skills add https://github.com/isizono/calm --skill rule-placement-isizono
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rule-placement
Source: https://github.com/isizono/calm/tree/main/skills/rule-placement
Command: npx skills add https://github.com/isizono/calm --skill rule-placement-isizono

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? When a user ruling or request contains rules that should change an AI agent's future behavior, it is unclear where to place them: habits, rules files, tag-notes, pins, feedback entries, precedent decisions, skills, docstrings, or CLAUDE.md. Misplaced rules either never fire or waste expensive always-on context. This Skill evaluates each rule and assigns it to the right delivery channel. ## Core Features & Use Cases - Rule decomposition: Splits a natural-language ruling into independently revocable rules before evaluation. - Four-axis evaluation: Scores each rule on trigger, type, risk, and revision likelihood, then maps the result through a lookup table to a candidate channel. - Approval-aware execution: Runs no-approval placements (feedback entries, tag-notes, pins, auto-memory) immediately and presents only approval-required placements (rules, habits, precedent decisions) as a plan table for user confirmation. - Use Case: After a meta-ask ruling like "always run tests before committing, and check the deploy log when touching infra files", the Skill splits it into two rules, places the first in habits and the second in path-scoped rules, then closes the loop with triage_ask promote. ## Quick Start Evaluate this ruling and decide which delivery channel each generalized rule should be placed in before promoting the meta-ask.

Frequently Asked Questions about rule-placement

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide where to store a new behavior rule for an AI agent?▼

Evaluate the rule on four axes: trigger (when it is needed), type (norm, conditional criterion, procedure, or fact), risk (miss cost times unawareness), and revision likelihood. The trigger-by-type lookup table then yields a candidate channel such as habits, rules, tag-notes, pins, or precedent decisions.

What is the difference between habits, rules, and tag-notes for rule placement?▼

Habits and rules are always-on push channels injected at session start, with habits limited to unconditional norms under 100 characters. Tag-notes fire on the first tool call touching a tag within a session, making them cheaper for context-scoped rules.

When should a rule go into a precedent decision instead of a push channel?▼

Conditional judgment criteria with self-aware doubt as the trigger belong in precedent decisions, which are pulled via search or bundled with similar asks. Low-to-medium risk rules default to precedent decisions because misplacement harm is minimal and later promotion is possible.

Can high-risk rules rely on pull-based delivery channels?▼

No. Rules with high miss cost and unawareness must be promoted to push channels even if no trigger can be named, because pull channels like intelligently habits have measured cases of never being retrieved. Use tag-notes, pins, or always-on push instead.

Why are skills and docstrings rejected for rules likely to be revised?▼

Skills, docstrings, and server instructions require PR and deployment to change, so rules with high revision likelihood cannot be placed there. Operate them via tag-notes or precedent decisions until revisions stop, then migrate through the PR path and retract the interim placement.