Claude Skills by badgids

Coordinate local-first generation on machines where Pi's local LLM and ComfyUI models cannot safely occupy GPU or system memory at the same time, using a fully precompiled ComfyUI batch and a model-free runtime handoff.

Revise a completed or partial book at manuscript, arc, chapter, scene, continuity, and prose levels while preserving chapter IDs and updating dependent plans when structure changes.

Review and revise a screenplay for structure, scene purpose, filmability, staging, dialogue, runtime, and page economy while keeping its scene manifest synchronized.

Diagnose and revise an existing prose story in separate structure, scene, character, continuity, and prose passes, fixing root causes before sentence polish.

Convert beats into executable scenes with stable scene IDs, location, time, characters, goal, conflict, turn, outcome, and causal handoff to the next scene.

Plan a film score as dramatic cues with entry, exit, duration, energy curve, instrumentation, tempo, key, motif rules, hit points, silence, and model targets.

Manage feature-film SEQ-### production boundaries so long films are planned, generated, reviewed, and approved in bounded sequence units.

Create or refresh a compact durable Story-Film handoff so a fresh Pi or agent session can resume the exact creative or production task without relying on chat memory.

Compile screenplay lines, performance blocking, shot briefs, scene state, and dialogue timing into a portable machine-readable shooting script with positions, moves, actions, camera coverage, exact dialogue, and timing sources.

Design motivated shots from a screenplay scene and director book, giving every shot a purpose, framing, camera behavior, action, continuity state, duration, and sound plan.

Compile approved shot briefs into a practical CSV shot list with source line IDs, setups, framing, angle, movement, action, audio, duration, continuity, and production notes.

Create an editable Shotcut .mlt project from canonical editorial state using MLT XML plus Shotcut project, bin, track, audio-mix, effect, transition, subtitle, and portable resource metadata.

Plan a coordinated social-media release campaign with verified facts, audience segments, content pillars, platforms, placements, aspect ratios, deliverable matrix, spoiler rules, accessibility, and a relative or dated campaign calendar.

Write coordinated campaign captions, hooks, calls to action, accessibility text, hashtags, keywords, and platform variants using verified release facts and an explicit no-invented-claims rule.

Build short social-video edits from approved film, trailer, and pickup media using stable source IDs, destination-specific timing, hooks, subtitles, title cards, and executable timelines.

Create aspect-specific social derivatives using approved focus points or fit rules for 16:9, 9:16, 1:1, 4:5, or custom dimensions without arbitrarily cropping required story information.

Design ambience, Foley, hard effects, transitions, perspective, silence, and generated SFX cues tied to scenes and shots, then route cues to Stable Audio 3 or video-native audio when appropriate.

Adapt music, stem, ambience, Foley, sound effect, audio-to-audio, inpainting, or continuation cues into Stable Audio 3 prompts with source, action, production character, duration, and model mode.

Design or repair a story core: logline, dramatic question, protagonist goal, opposition, stakes, change, ending logic, and a causal architecture appropriate to the requested form.

Turn an idea, conversation, synopsis, or assignment into a compact creative brief with format, premise, audience, dramatic question, constraints, must-haves, and open decisions.

Run and interpret the story-film suite's weak-model, continuity, adapter, style, and prompt-injection evaluation cases. Use when developing or regression-testing the skill package itself.

Self-contained Story-Film Skills bundle for story, book, screenplay, image, audio, video, film, postproduction, and release work. Use this entry point when Story-Film Skills was installed with npx skills.

Adapt MiniMax H3 prompts for stylized 3D narrative animation with locked character identity, environment continuity, readable acting, per-shot timing, physical action, camera, and audio continuity.

Adapt score or song cues into ACE-Step 1.5 XL inputs using task type, caption, lyrics or instrumental structure, BPM, key, time signature, and temporal performance directions.

Map an existing story, novel, treatment, or source narrative into a screen adaptation plan that identifies what becomes action, dialogue, visual information, compression, combination, or omission.

Register and approve generated or imported media across reference, voice, dialogue, music, SFX, video, master, trailer, social, artwork, and release groups with one durable primary/alternate/reject state model.

Simulate a defined reader or viewer encountering prose, screenplay, storyboard, animatic, or rough-cut material for the first time, reporting felt attention, questions, character legibility, clarity, and payoff without turning into craft critique.

Build and render a synchronized full-program audio master from approved dialogue, voiceover, ambience, Foley, SFX, score, and music using an explicit project-relative mix manifest.

Recover only failed, unfinished, and invalidated downstream ComfyUI jobs while preserving valid completed generation work.

Turn story architecture into a causal beat sheet whose beats state purpose, cause, action, turn, and consequence without padding to a formulaic beat count.

Plan a novel, novella, or nonfiction narrative book as a durable chapter architecture with POV rules, part structure, chapter jobs, arcs, reveals, continuity state, and target length.

Adapt MiniMax H3 prompts for short brand, product, app, site, or campaign promos using verified brand facts and assets, product-specific proof, precise beats, readable copy, and controlled motion.

Define and enforce durable campaign voice, terminology, naming, CTA, spoiler, accessibility, and prohibited-claim rules so trailers, social copy, press materials, and promotional art stay recognizably consistent.

Render and verify the planned social campaign deliverables across aspect ratios and durations, then reconcile masters, copy, artwork, captions, subtitles, and QC into a complete campaign delivery set.

Create a compact character bible focused on behavior under pressure, goals, contradictions, relationships, speech, appearance, wardrobe, and continuity facts that affect story or production.

Pressure-test a character by speaking or reasoning strictly from that character's current voice, goals, relationships, emotional pressure, and knowledge boundary without creating canon by accident.

Adapt MiniMax H3 prompts for two-player co-op game menu and opening animations with locked player identity, stable UI hierarchy, exact game/player text, coordinated palette, and timed menu interaction.

Operate a user-selected Comfy API v2 endpoint such as comfy-api-proxy using durable jobs, poll-first state, UUID assets, API-format workflow submission, targeted cancellation, safe asset retrieval, optional bearer auth, and idempotent submission practices.

Stage ComfyUI workflow input media through server uploads, use returned server names instead of guessed paths, extract all file-shaped outputs from history, download outputs safely, and keep project-relative media records.

Audit reference-to-workflow bindings so prompt ordinals, REF/MEDIA identities, staged files, hashes, and actual graph nodes cannot silently drift.

Use the official comfy-cli safely for ComfyUI workspace discovery, setup, launch and stop, structured JSON automation, workflow execution including UI-format conversion, jobs, uploads, downloads, nodes, models, templates, memory, updates, and approved dependency changes.

Inspect a live ComfyUI server for version, hardware, feature flags, queue depth, installed node schemas, model categories, and model filenames before selecting or validating a workflow.

Build a standalone, portable ComfyUI-ready generation manifest containing canon links, references, briefs, prompt files, selected workflow identities, input requirements, output IDs, stale scope, and unresolved requirements without hardcoded machine details.

Use Story-Film's Pi-native story_comfy tool and managed official comfy-mcp runtime to inspect, validate, run, monitor, and collect ComfyUI work with live discovery and approval gates.

Compile image, audio, and video generation into a complete deterministic ComfyUI batch that can execute without any LLM involvement after handoff.

Execute API-format workflows on ComfyUI, keep prompt IDs, poll authoritative history, inspect queue state, cancel selected runs, free ComfyUI memory, capture errors, and record outputs for story-film production.

Diagnose ComfyUI connection failures, invalid workflows, unknown node classes, missing required inputs, missing models, queue and history problems, generation errors, VRAM pressure, remote-target confusion, and dependency issues without random installs or broad destructive changes.

Select, preserve, inspect, validate, patch, convert, and promote complete ComfyUI workflows from Story-Film's extension-local comfyui_workflows library.

Understand and operate ComfyUI for story-film production through Story-Film's extension-local workflow library, managed official comfy-mcp/comfy-cli control runtime, live node/model discovery, deterministic validation, generation execution, API v2, offline batching, resource handoff, and safe troubleshooting.

Adapt approved film, trailer, transcript, press, and campaign material into destination-specific content while preserving source lineage, verified claims, spoiler policy, brand voice, and the central promise.