Professional skill for output compression in coding and agent workflows.
I don't like babysitting cavemen. I want to talk to professionals and go an adventures, while mission control has my back.
Most "caveman" skills reduce token use, but add new problems:
- They cut readability too aggressively.
- Efficency is just a byproduct, not the core design goal.
- They sound unprofessional → Costly context switching if output is used in commit messages, PR comments, or client-facing text.
Spaceman keeps the speed gains without those costs.
Spaceman uses a simple metaphor:
- User = Spacefarer under comms delay.
- Assistant = Mission Control with high context and responsibility.
- Constraint = Every extra word delays the next correct action (this is dangerous).
This framing creates a stable optimization target: maximize information-per-token and minimize time-to-action while preserving professional quality.
- Debugging and root-cause analysis (you need to justify the extra tokens for reading the skill)
- Implementation plans
- Code- or security reviews
- Architecture and design, investigating trade-offs and comparisons
- Interactive learning and working with knowledge bases
Use the skill here in Github Copilot, OpenCode, Claude Coder or any agent workflow that supports skills:
Activate the skill with a /spaceman command or ask for:
- "spaceman mode"
- "mission control mode"
- "compress output"