plan-writer-fastlisted

# Plan writer — fast path The Obelus desktop spawns this skill for writer-mode bundles ("draft these bullets into prose", "tighten this section"). The user reviews every diff in the Obelus diff-review UI before applying, so the value here is **speed**: one LLM turn that reads the bundle, reads the whole paper source, plans a coherent edit set, writes the plan, and ends. This skill **does not** run the structural review the `apply-revision` → `plan-fix` path does. No subagent stress-test, no impact sweep, no coherence sweep, no quality sweep. If the user wants any of those, they pick **Rigorous** in the UI and the desktop spawns `apply-revision` instead. This skill **does not** edit any source file. It emits a plan; the user runs `apply-fix` (or clicks Apply in the desktop) when they are ready. ## Pacing rule — emit phase markers BEFORE deep reasoning Each `[obelus:phase] <name>` marker (listed below) must be emitted **on the assistant's first text output of that phase, before any large thinking block or any tool call**. Do not pre-think the entire phase before emitting the marker. Tool calls are cheap; thinking blocks are not. A 30k-character thinking block before the first `[obelus:phase]` of a phase is the single most expensive failure mode of this skill. ## File output contract — non-negotiable Emit one artefact per run, under `$OBELUS_WORKSPACE_DIR/`, stamped with a compact UTC timestamp generated once at the start of the run (`YYYYMMDD-HHmmss`, e.g. `20260423-1430