Derive per-character mission formulas and grade elicited replies against predicates.
npx skills add https://github.com/mrrts/WorldThreads --skill derive-and-test
Please help me install this Agent Skill. Skill: derive-and-test Source: https://github.com/mrrts/WorldThreads/tree/main/.agents/skills/derive-and-test Command: npx skills add https://github.com/mrrts/WorldThreads --skill derive-and-test
Derives a per-character/per-world MISSION FORMULA and executes a targeted elicitation against it, producing a reusable substrate update and a structured return reading.
Run derive-and-test for a specific character and world to derive the formula, elicit a test prompt, and generate a report.
High-intent search queries and answers about installing and using this skill.
Character behavior prompts are tested by generating a per-character, per-world mission formula and running a single elicitation against it. The reply is then graded against the derived predicates to produce a structured behavior report.
A living substrate update is a return-reading style report generated after grading a test elicitation. It refreshes the character's living document with the resulting behavioral data to maintain a continuous world state.
Deriving a mission formula involves generating a triadic instantiation that grounds character behavior within a specific world. This creates a reusable formula used to execute and grade targeted prompt elicitations.
Yes, an authorization-based chooser exposes forecasted next steps and their associated costs. This allows you to authorize follow-on work and refine the derivation using a running trial of prior runs.
No external dependencies or components are required to run derive-and-test. It operates independently to derive the mission formula, elicit test prompts, and update the living substrate.
A per-character test loop is necessary when you need to ground behavior in a specific world instantiation and maintain a living substrate. It provides triadic derivation and predicate grading rather than isolated prompt evaluation.
jeffstricGenerate and manage character reference images in bulk using world style and character data.
marktodaInventory codebase specifications, invariants, gotchas, linters, and CI checks.
joeloverbeckAudit branching-story bundles for structural integrity, content quality, and narrative consistency.
Ikram-AlamEvaluate agent outputs against rubrics and ground truth with multi-dimensional scoring.
