RPCS-1 Agent Tuner & Translation Bridge

RPCS-1 turns an agent's operating conditions into derived runtime settings (temperature, context strategy, failure-mode read), and its Translation Bridge…

От сообществаРаботаетБез входаRU / СНГБесплатноТолько чтение

Что умеет

Recommend Agent Configuration: Diagnose why a deployed AI agent may fail. Returns suggested runtime parameters, a context strategy, a predicted regime, reasoning, warnings, and a next test. Determinis
Interpret: Detect ambiguity in user messages. Returns AR level, confidence, candidate interpretations, clarifying questions, and a suggested next step.
Normalize: Clean up fragmented human input with ellipses and run-on thoughts into coherent prose.

Какие данные видит

Нужен ли аккаунт

Не нужен: сервер работает без входа

RPCS-1 turns an agent's operating conditions into derived runtime settings (temperature, context strategy, failure-mode read), and its Translation Bridge calibrates a per-user receiver profile so every reply is rendered for that user's communication style. 7 tools, no API key.

Список инструментов сервера (7)

Технические названия из tools/list. Нужны только разработчикам.

recommend_agent_configuration	Diagnose why a deployed AI agent may fail. Returns suggested runtime parameters, a context strategy, a predicted regime, reasoning, warnings, and a next test. Deterministic, stateless, and read-only. Optionally pass target_model to attach measured per-model receiver posture (evidence-graded).
interpret	Detect ambiguity in user messages. Returns AR level, confidence, candidate interpretations, clarifying questions, and a suggested next step.
normalize	Clean up fragmented human input with ellipses and run-on thoughts into coherent prose.
rewrite	Return style-specific rewrite instructions for technical, plain, socially gentle, concise, detailed, or direct communication.
calibrate_profile	Build a continuous communication-preference profile from five behavioral forced-choice answers. Omit answers to receive the questions. The profile is not a diagnosis or category label, and nothing is stored server-side.
prepare_prompt	Recover likely intent from a raw message before acting. Returns a canonical translation, ambiguity level, and whether to clarify or commit; an optional receiver profile shapes that choice.
render_reply	Return deterministic rendering instructions for a draft reply and a receiver profile. Apply the returned structure, warmth, explicitness, revision, and ambiguity guidance before answering.