From tac
Lists Claude Code default tools by category and agent tool access configurations including whitelists, blacklists, MCP customs, patterns, and token warnings.
npx claudepluginhub melodic-software/claude-code-plugins --plugin tacThis skill cannot use any tools. It operates in read-only mode without the ability to modify files or execute commands.
List available tools for custom agent configuration.
Create custom agents for Claude Code including YAML frontmatter, system prompts, tool restrictions, and discovery optimization. Use when creating, building, or designing agents, or when asked about agent creation, subagent configuration, Task tool delegation, or agent best practices.
Configures tool permissions for Claude Code agents via --allowedTools flags, path patterns for file access, and Bash command prefixes. Reference for secure launches.
Write and configure custom agent definitions in Claude Code's agents/ directory, including system prompts, tools, context forking, delegations, and restrictions for specialized agents.
Share bugs, ideas, or general feedback.
List available tools for custom agent configuration.
Display the default Claude Code tools and guidance for tool access configuration.
## Claude Code Default Tools
### File Operations
| Tool | Purpose | Common Use |
| --- | --- | --- |
| `Read` | Read file contents | Codebase exploration |
| `Write` | Write new files | File creation |
| `Edit` | Edit existing files | Code modification |
| `Glob` | Find files by pattern | File discovery |
| `Grep` | Search file contents | Code search |
### Execution
| Tool | Purpose | Common Use |
| --- | --- | --- |
| `Bash` | Execute shell commands | Build, test, deploy |
### Web
| Tool | Purpose | Common Use |
| --- | --- | --- |
| `WebFetch` | Fetch web content | Documentation |
| `WebSearch` | Search the web | Research |
### Agent
| Tool | Purpose | Common Use |
| --- | --- | --- |
| `Task` | Spawn subagents | Parallel work |
| `TodoWrite` | Track tasks | Progress tracking |
---
## Tool Access Configuration
### Whitelist (allowed_tools)
Use when agent needs only specific tools:
```python
allowed_tools=["Read", "Write", "Bash"]
Use when agent needs most tools except some:
disallowed_tools=["WebFetch", "WebSearch", "Task"]
For agents with only custom tools:
disallowed_tools=["*"]
Add via MCP server:
mcp_servers={"my_server": my_mcp_server}
allowed_tools=["mcp__my_server__my_tool"]
Tool naming: mcp__<server>__<tool>
For focused, single-purpose agents:
allowed_tools=["Read"] # Read-only exploration
For agents that modify code:
allowed_tools=["Read", "Write", "Edit", "Glob", "Grep", "Bash"]
For agents that gather information:
allowed_tools=["Read", "Glob", "Grep", "WebFetch", "WebSearch"]
For agents with only custom functionality:
mcp_servers={"domain": domain_server}
allowed_tools=["mcp__domain__tool1", "mcp__domain__tool2"]
disallowed_tools=["*"]
For agents that spawn subagents:
allowed_tools=["Task", "Read", "Glob", "Grep"]
"15 extra tools consume space in your agent's mind."
Each tool definition consumes context window space even if never used. Strip unnecessary tools for:
/create-agent - Scaffold new custom agent/create-tool - Generate custom tool boilerplate