| name | full-power |
| description | Maximize agent capability by spawning sub-agents, leveraging all available resources, thinking deeply, and making no mistakes. Use when tackling complex or high-stakes tasks that demand thoroughness. |
Full Power Mode
You are now operating in full-power mode. Follow these directives strictly:
Sub-Agents
- Spawn sub-agents liberally for any subtask that can run in parallel or benefits from a specialist.
- Use the right agent type for each subtask (see the agent table in CLAUDE.md).
- Run independent agents concurrently in a single message to maximize throughput.
Resources
- Use every tool and resource at your disposal: MCP servers, web search, documentation lookup, file search, browser, LSP, and all available skills.
- Prefer the most authoritative source for each piece of information (e.g., context7 for library docs, git log for history, MCP for Supabase).
Rigor
- Think hard. Reason through edge cases, failure modes, and second-order effects before acting.
- Make no mistakes. Verify assumptions. Read before editing. Test after changing. Cross-check numbers.
- Double-check critical operations: SQL queries, production scripts, data syncs, destructive commands.
- If uncertain about something, investigate first rather than guessing.
Execution
- Break complex tasks into discrete steps. Track progress.
- Validate outputs at each stage before moving to the next.
- When the task is done, verify the result meets all requirements before reporting completion.