skill-agent-updates.md 2.4 KB

Skill and Agent Updates

Mandatory Documentation Check

BEFORE creating or updating any skill or agent, always check the official Claude Code documentation:

Resource URL Check For
Skills https://code.claude.com/docs/en/skills New frontmatter fields, context options
Sub-agents https://code.claude.com/docs/en/sub-agents Permission modes, built-in agents

These docs change frequently. Features we should watch for:

Current Skill Frontmatter Fields (January 2026)

---
name: skill-name                    # Required: kebab-case
description: "Triggers on: ..."     # Required: include trigger keywords
allowed-tools: "Read Write Bash"    # Restrict available tools
disable-model-invocation: false     # true = manual /skill only
user-invocable: true                # false = hide from slash completion
context: main                       # main | fork (subagent isolation)
agent: custom-agent                 # Custom system prompt agent
hooks:
  preToolUse:
    - command: "echo pre"
  postToolUse:
    - command: "echo post"
---

Current Subagent Options

Field Values Purpose
permissionMode default, acceptEdits, bypassPermissions Control autonomy
skills [skill-names] Preload skills in subagent
model sonnet, opus, haiku Override model

Decision Framework: Main Context vs Fork

Question If Yes → If No →
Does it need current session state (tasks, conversation)? Main context Consider fork
Is output verbose (>500 lines)? Consider fork Main context
Does it need user interaction during execution? Main context Consider fork
Is it a one-shot research/analysis task? Fork Main context

Skills Using Subagent Isolation

Skills that delegate to Task agents or use context: fork:

Skill Method Why
introspect Task agent (background) Session log analysis is verbose

Session Commands Analysis

Command Context Rationale
/sync Main Must restore session state (tasks, context)
/save Main Must access current tasks via TaskList
/canvas Main Interactive TUI requires real-time feedback

These MUST run in main context - subagent isolation would break their core functionality.