output-compress
Tiered output compression (caveman-derived): an explicitly opt-in, token-saving rewrite mode with a never-compress whitelist, model-tier compression caps, and a deterministic fidelity gate (no LLM self-judgment). Use when the user types output-compress, /compress, "compress lite|full|ultra", or asks to shorten/condense internal or scratch output. Do NOT use for: the final user-facing response language, safety/irreversible-action confirmations, contract fields (Goal/Non-goals/Done-when/Return), audit or review findings that must stay verbatim, or as a default/always-on behavior.
Source facts
- Repository
- zeuikli/output-compress
- Last source activity
- July 17, 2026 at 02:04
- Detected SKILL.md language
- English
- Stars
- 4
- Forks
- 0
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Showing SKILL.md
| name | output-compress |
|---|---|
| description | Tiered output compression (caveman-derived): an explicitly opt-in, token-saving rewrite mode with a never-compress whitelist, model-tier compression caps, and a deterministic fidelity gate (no LLM self-judgment). Use when the user types output-compress, /compress, "compress lite|full|ultra", or asks to shorten/condense internal or scratch output. Do NOT use for: the final user-facing response language, safety/irreversible-action confirmations, contract fields (Goal/Non-goals/Done-when/Return), audit or review findings that must stay verbatim, or as a default/always-on behavior. |