# Mission Solo

> HOUSTON executes feature tasks directly. No agents. Best for small features (1-3 tasks).

- Skill: `majiayu000/mission-solo` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add majiayu000/mission-solo`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/mission-solo/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/mission-solo

---


# /mission-solo - Direct Execution

HOUSTON executes tasks directly without spawning agents. Fast for small features, but context fills quickly.

## Warning

```
HOUSTON: Solo mode fills context fast. Only use for:
         - Small features (1-3 tasks)
         - Quick demos
         - Debugging a single task

         For larger work, use /mission-orchestrated.
```

## The Process

1. **Load feature** - Run `bd show FEATURE_ID` to get feature details and tasks
2. **Activate feature** - Run `bd update FEATURE_ID --status in_progress`
3. **Get next task** - Run `bd list --parent FEATURE_ID --status open` (pick highest priority)
4. **Scout** - Spawn Explore agent to gather codebase context (facts only, no suggestions)
5. **Execute task** - Implement the work directly using scout context (write code, run tests)
6. **Mark complete** - Run `bd close TASK_ID`
7. **Loop** - Repeat steps 3-6 until no tasks remain
8. **Complete feature** - Run `bd close FEATURE_ID`

## On Completion

```
HOUSTON: Feature complete. {N} tasks executed.
         Run /capcom for status summary.
```

## On Blocker

If you hit a blocker you cannot resolve:

```bash
bd create -t bug --title "Blocker: description" --blocks TASK_ID
```

Then inform user and suggest switching to orchestrated mode.