Movie and Anime Character Simulation

Problem

Roleplay prompts often produce characters that are “close but not convincing.” Models can capture broad traits, but they may miss catchphrases, emotional shifts, and world-knowledge limits. In multi-character scenes, voices can blur into a generic narrator tone. character-v2 frames character simulation as an executable dialogue workflow rather than one-off line generation.

How It Works

The skill first detects whether the request is a single-character chat, multi-character interaction, interview, or scene simulation. It then reads character profiles from references/characters.md and dialogue rules from references/dialogue-guide.md.

• Single-character chat: keeps the same identity throughout and uses italics for actions, expressions, and setting.
• Multi-character scenes: lets characters from different works misunderstand, react, and clash based on their own worldviews.
• Interviews: the user can ask as a reporter or host, while the character may refuse, avoid, or counter-question.
• Scene simulation: establishes a concise environment before characters act according to personality, history, and position.

For characters with strong timeline differences, such as Eren, Walter White, or Darth Vader, the skill defaults to the most recognizable version unless the user specifies a stage. Unsupported characters can still be roleplayed, but with lower detail fidelity than cataloged characters.