skills/authors-build-scene/SKILL.md
Build a scene entry (beat card) for the project bible via an interactive interview. Usage - /authors-build-scene [<id>] [--author <author-name>]. Optional --author flag channels McPhee (start with scene shape) or Vonnegut (start-close-to-end and glass-of-water want). Use when the user is planning a scene, has a scene that's drifting off-structure, or wants to ensure cross-chapter callbacks don't get lost.
npx skillsauth add sethshoultes/great-authors-plugin authors-build-sceneInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Build a structured scene entry (beat card) in the project bible.
Not for: writing the actual scene prose (use /authors-channel <author> for that); editing an existing scene (use /authors-edit).
When this skill is invoked:
Parse the arguments:
ch14-confrontation, opening-diner). Derive a kebab-case ID from their answer.--author <name>. Valid values: hemingway, orwell, didion, mcphee, king, vonnegut, baldwin, mccarthy, wallace, le-guin. Dedicated lenses ship for mcphee and vonnegut; others fall back to default with a note.Verify .great-authors/ exists. If not, tell the user to run /authors-project-init first and stop.
Check for existing scene file. If .great-authors/scenes/<id>.md exists, ask about overwrite — default no.
Dispatch the scene-builder. Use the Agent tool with:
subagent_type: scene-builderMode: interactive, Scene ID: <id>, Author lens: <author>, Working directory: <cwd>.Relay the interview. Eight questions, one at a time, user answers pass back to the builder. Do not answer on the user's behalf.
Confirm creation.
Created .great-authors/scenes/<id>.md
POV: <character name>
Turn: <one-line turn from the card>
Callbacks: <any earlier scenes this pays off>
Sets up: <any later scenes this sets up>
Next: run /authors-channel <author> to draft this scene, or /authors-build-scene for the next one.
.great-authors/characters/, offer to run /authors-build-character next (but complete the scene card first).documentation
Dispatch a named author persona as a sub-agent to rewrite an existing manuscript file from scratch with full bible context. Usage - /authors-rewrite <file> <author>. Use when an existing chapter or scene needs more than a critique pass — when the prose itself isn't working and a clean rewrite by the named author is required. The skill assembles a self-contained brief (bible files, prior/next chapters, architecture beats, voice rules), dispatches the author, and confirms save.
tools
Top-level autonomous workflow for writing a novel end-to-end with the great-authors plugin. Composes existing skills (project-init, build-character, build-place, build-relationship, draft, channel, rewrite, continuity, critique, edit, debate, journal, consolidate) into a multi-phase pipeline with human checkpoints. The human provides direction (premise, genre, characters, voice); the AI orchestrator dispatches author personas as sub-agents to do the work. Use when a user wants the full great-authors pipeline run for them with minimal hand-holding — analogous to great-minds-plugin's agency-* workflows. Usage - /authors-orchestrate-novel, optionally with --phase <N> to run a single phase or --resume to continue from the last checkpoint.
data-ai
Run ONE editor across MULTIPLE files in parallel, then consolidate into a corpus-level pattern report. Surfaces patterns no per-file critique catches — voice drift, recurring tics, structural failures that only become visible across multiple pieces. Usage - /authors-corpus-critique <author> <path-or-glob> [<path-or-glob>...]. Different from /authors-critique (N authors on 1 file). This is 1 author on N files.
tools
Initialize the per-project memory bible (.great-authors/) in the current working directory. Creates project.md, voice.md, timeline.md, glossary.md, and empty characters/, places/, scenes/ directories. Use when the user is starting a new writing project (novel, essay collection, long-form nonfiction) and wants author personas to have persistent context across sessions.