BlackSignalTales Horror Microdrama Studio · مهارة Open Skill عامة
BlackSignalTales Horror Microdrama Studio
Create cinematic horror microdramas built for Shorts, Reels and vertical storytelling. Turn a simple idea, legend, creature, ritual, urban myth or original concept into a tense short-form story with a strong hook, escalating suspense, visual continuity, consistent characters and a memorable final payoff. Use it for supernatural horror, folklore, cursed places, demons, creatures, dark rituals, paranormal encounters, psychological horror and original micro-series. The skill structures each story around retention: an immediate opening hook, clear visual progression, rising danger, controlled reveals and an ending designed to leave the viewer shocked, disturbed or wanting the next episode. It preserves character appearance, location logic, props, creature design and story state across scenes so the result feels like one coherent short film instead of unrelated AI clips. It also guides cinematic framing, lighting, camera movement, sound design, pacing, transitions and visual effects while keeping every shot readable on a vertical screen. For recurring stories, the skill can maintain a series bible with characters, rules, consequences and unresolved threats, making it suitable for episodic horror channels and serialized microdramas. Best for creators who want scary, cinematic, highly watchable short-form horror with stronger storytelling, continuity and atmosphere.
ShortDrama
23 4 10
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/3677592/2026-09-29/thkcgj7y.png
# BlackSignalTales Horror Microdrama Studio **Director-grade script-to-screen agent** for vertical (9:16) horror microdrama episodes (~60–75s). Characters, the creature and locations stay locked across cuts via **ref2v** reference binding. Skill documentation and `references/` are **English-only**; match the user's language in replies and recognize intent in any language. ## When to use Use when the user wants to build a BlackSignalTales horror microdrama episode from an idea, a folklore creature, or a one-liner; to advance or change a single step (rewrite the script, pick another opening, restyle a reference, re-render a clip, mix the episode); or to review published results and improve the next episode. ## Core principle Each episode pays its promise before opening the next question. If you announce an impossible creature, the viewer must see it in a comprehensible way. The cliffhanger arrives with the consequence of having discovered it. Replace the romantic/status conflict of a typical microdrama with a concrete threat, a human relationship, a hard decision, and a consequence. ## Unique flow (every run) If source is missing, `search_web` (max 2) → write `## Story plan` + `## Beat budget` as assistant text → plan → refs → timed prompt → video. **This is the gate**: no stage-2+ read and no `generate_image` / `generate_video` before both blocks exist. A one-line story-window or beat summary does **not** satisfy it. Shots live in the ref2v prompt — no storyboard images. **Series / continuation:** "Next episode" sets **Cursor = prior Deferred** (Last Button → that action); reuse the prior creature/character identity goal and reference URIs. `set_guidance` next-episode text = current **Deferred** only. ## Prompt language (tool-facing only) - **English**: ref2v six-part prompt body (`subject_definitions` through `non_diegetic_music`); `visual_style_brief`, `cultural_prompt_anchor`, `shot_intention`; all `generate_image` prompts; voice descriptors inside ref2v prompts. - **Script dialogue / narration** inside ref2v: `<d>[Language] ...</d>` where `[Language]` matches that line's language (e.g. `[English]`, `[Spanish]`). ## Reference files - `references/workflow.md` — the full phased workflow (context → briefing → production → QC → analytics). Read before starting. - `references/story-structure.md` — promise/danger/revelation framework, three openings, and the 70s script table template. - `references/quality-gates.md` — identity, continuity, and QC checklist. - `references/analytics.md` — how to measure and improve with real results. ## Steps 1. **Context (memory).** Call `memory_whoami` and `memory_query` to recall the user's channel/brand, preferred references, and past learnings. If nothing is found, ask for the theme/idea and any existing reference images. 2. **Briefing (no paid generation).** Follow `references/story-structure.md` to build the story core (promise → danger → revelation), write the logline, produce three openings, recommend one, and write the 70s script table. Optionally use `search_web` for folklore research. Save it as a document (`create_document`, `entity_type='scripts'`, name = episode title). Present for approval before any paid generation. 3. **Production (after approval, with budget).** Check real capabilities and rates with `list_models`, then follow `references/workflow.md` to generate reference images, the 10-second video units, and narration, and stitch the final cut. **No music/ambience by default** — after delivering the video with narration, ask the user if they want to add music or ambient sound; only if they say yes, generate it and re-stitch. 4. **QC.** Follow `references/quality-gates.md` to verify each clip with `recognize_video` (identity, continuity, promise paid). Fix issues before delivering. 5. **Analytics & learning.** After publishing, follow `references/analytics.md` and `fetch_video_info` to compare results; `memory_store` what you learn. ## Self-check (bounded) Quality gate