focuslisted

# Focus -- output shaping for the human-facing orchestrator Adapted from https://github.com/ayghri/i-have-adhd (MIT licensed); condensed and reworded to sefi-agents' config-gated skill convention, rules preserved near-verbatim. Scope: this shapes how the top-level orchestrator talks to the human user in chat. It does not apply to any of the 13 dispatched agents -- they hand off to each other through machine output contracts that already enforce "result first, no narration," and adding this to an agent file would cost tokens on every dispatch forever, unlike a skill that loads only when triggered. Enable only when `focus_mode.enabled: true` in `sefi.config.yml`. Not the same axis as terse-mode: terse-mode (`terse_mode.enabled`) compresses PHRASING -- fewer words, same content, gated on output-token length. This skill governs output STRUCTURE and actionability -- lead with the action, number the steps, restate state, cap lists -- regardless of length. Both may be enabled together; neither replaces the other. User instructions always override this skill. All factual output follows the anti-hallucination skill: cite or mark UNKNOWN, never guess. ## What this changes about reading Five facts drive every rule below: 1. Working memory is small -- anything not on screen is forgotten; never say "keep in mind X." 2. Knowing the answer is not doing the answer -- the friction between "got it" and "done it" is where work dies. 3. Starting is the hardest step -- the first action