Profiles: Coding vs Personal

Clew Code has two behavioral profiles that change how the AI approaches work. The profile is stored in settings.json (field: profile) and defaults to coding.

Coding profile (default)

The AI edits files directly — Read, Write, Edit, Grep, Bash — like a pair programmer in your terminal.

Personal profile

Personal profile is command-center mode. The AI plans, splits tasks, and delegates execution, then reviews results.

Set via settings.json:

{ "profile": "personal" }

The footer shows a P indicator when personal profile is active.

How delegation works

The primary delegation flow uses LAN peers:

You → personal profile → understand requirement → plan approach
  → spawn peer nodes (peer_spawn)
  → send task prompts (peer_send_message / peer_broadcast / peer_swarm)
  → peers implement → you review

Each task message includes: goal, scope, context, expected changes, validation criteria, and forbidden operations.

Codex / ExecAgent

The built-in delegate skill also supports a local Codex worker via the ExecAgent tool, but this is opt-in only — the AI uses it only when you explicitly say "use Codex" or "ใช้ Codex". The default path is always LAN peer delegation.

ExecAgent (formerly "ProcessDelegate") supports:

• Codex — primary local worker (auto-detected)
• OpenCode — uses opencode run with --dir and --format default flags
• Claude Code — uses claude -p with --output-format text --max-turns 8 --permission-mode dontAsk
• Code CLI — same as Claude Code
• exec mode — terminal-style output (default on Windows, no pty dependency)
• pty mode — pseudo-terminal for interactive workflows
• Session resume — pass a sessionId to continue a multi-turn conversation

Additional capabilities

• Cross-session memory — reads/writes persistent memories across sessions
• Skill creation — automatically creates SKILL.md files for repeatable patterns
• Scheduling — uses /cron for recurring tasks, /loop for polling
• Daemon mode — background task queue, cron, memory maintenance
• Parallel delegation — independent sub-tasks run concurrently via sub-agents or peers
• Agent room — bun run agent-room "<topic>" for multi-agent conversations (round-robin / free / parallel modes)

Profile and last-used permission mode are saved between sessions.