no-prepost-user-input

Preserve LLM decisions on user input without deterministic pre- or post-processing.

3|Updated Mar 22, 2026

Search Tags:#llm #prompts #state-persistence #input-interpretation #prepost-processing #deterministic-filtering

Ready to run

npx skills add https://github.com/eamars/KazusaAIChatbot --skill no-prepost-user-input

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: no-prepost-user-input
Source: https://github.com/eamars/KazusaAIChatbot/tree/main/.agents/skills/no-prepost-user-input
Command: npx skills add https://github.com/eamars/KazusaAIChatbot --skill no-prepost-user-input

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Preserve the LLM's decisions about user input by preventing deterministic pre-processing or post-processing from altering acceptance, commitments, or routing of commands.

Core Features & Use Cases

Maintains responsibility for interpreting user commands within the LLM, avoiding local rule-based overrides.
Supports cognition preference handling, consolidator facts harvesting, and evaluator prompts deciding channels for results.
Helps ensure durable state persistence of accepted commands and commitments without semantic rewrites.

Quick Start

Ask the model to determine user-input interpretation without applying deterministic pre/post filters and persist the decisions as durable state.

Frequently Asked Questions about no-prepost-user-input

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I preserve LLM user input decisions during post-processing?▼

To preserve LLM user input decisions, you must decouple user-input interpretation from local deterministic post-processing. This ensures the LLM's original decisions regarding acceptance, commitments, or command routing are not altered by semantic rewrites or rule-based overrides.

What is deterministic pre-processing and post-processing in LLM prompts?▼

Deterministic pre-processing and post-processing in LLM prompts refers to local rule-based filters that alter user-input interpretation. This Skill prevents these deterministic filters from modifying the LLM's original semantic decisions, ensuring durable state persistence for accepted commands without semantic reinterpretation.

How do I persist LLM accepted commands and commitments as durable state?▼

You persist LLM accepted commands and commitments by using explicit prompts, schemas, and evaluation prompts to emit durable fields. This approach maintains state persistence by keeping the responsibility for interpreting user commands within the LLM rather than applying local semantic rewrites.

When should I decouple user-input interpretation from local post-processing?▼

You should decouple user-input interpretation from local post-processing when handling cognition preferences, consolidator facts harvesting, or evaluator prompts determining channels for results. This ensures the LLM's decisions are preserved without semantic reinterpretation altering the acceptance or routing of commands.

Does decoupling LLM interpretation require explicit schemas and evaluation prompts?▼

Yes, decoupling LLM interpretation requires explicit prompts, schemas, and evaluation prompts to emit durable fields. These components are necessary to ensure that accepted commands and commitments are persisted accurately without undergoing semantic reinterpretation from deterministic filtering.

Why does deterministic filtering alter LLM command routing and acceptance?▼

Deterministic filtering alters LLM command routing and acceptance because local rule-based overrides reinterpret the semantic meaning of user inputs. Decoupling user-input interpretation from these filters prevents the deterministic post-processing from modifying the LLM's original decisions and commitment persistence.