Build psychological profiles from interaction data and predict cognitive trajectories.
npx skills add https://github.com/plurigrid/asi --skill cognitive-surrogate
Please help me install this Agent Skill. Skill: cognitive-surrogate Source: https://github.com/plurigrid/asi/tree/main/skills/cognitive-surrogate Command: npx skills add https://github.com/plurigrid/asi --skill cognitive-surrogate
The Cognitive Surrogate Skill enables building high-fidelity psychological models from interaction data. It extracts values, predicts cognitive trajectories, and generates authentic responses that preserve the subject's voice with high fidelity.
Core Principle: A surrogate is a derivational continuation of cognition, not a surface imitation.
NEW (Langevin/Gibbs Integration): Predictions leverage a Gibbs distribution; confidence scores reflect mixing time and temperature parameters.
Initialize a seed, build a profile from your interaction corpus, train a predictor, and generate an authentic reply.
High-intent search queries and answers about installing and using this skill.
Psychological models extract values and predict cognitive trajectories from interaction patterns. Initialize a seed, build a profile from your dialogue corpus or interaction logs, train a predictor, and generate authentic replies that preserve the subject's voice with >90% fidelity.
A cognitive surrogate is a derivational continuation of cognition, not a surface imitation. It reconstructs psychological patterns from interaction data to generate predictions and responses grounded in the subject's actual reasoning patterns.
Yes. The Langevin/Gibbs integration enables predictions that leverage Gibbs distributions, with confidence scores reflecting mixing time and temperature parameters for more precise uncertainty quantification.
Validation is configurable and meets functional requirements such as >90% fidelity in predictions. Build your profile, train the predictor, then run validate-fidelity against held-out interaction data to measure prediction accuracy.
Built-in governance includes explicit subject consent, transparent disclosure when surrogate-generated content is used, boundaries on high-stakes decisions, and audit trails with seeds for reproducibility.
Yes. The Skill applies across dialogue corpora, interaction logs, and consented research contexts. It integrates data extraction pipelines via DuckDB and multi-interpreter support for flexible input handling.
Snowflake-LabsA concise summary of your roleplay character: name,</name> personality, and/or behavior, formatted as a single string sentence. Typically不超过50字，此处省略
SignalPilot-LabsEnforce dbt write patterns for column naming and type preservation.
plurigridTrain learning agents on interaction sequences to discover behavioral patterns.
snoodleboot-ioExplains machine learning model predictions using SHAP, PDP, ICE, and counterfactual methods.
