本文へスキップ
TakoAPI
スキル一覧に戻る

sera-lexicon

wentinkjason 作

Coding Agents & IDEs
1.3kダウンロード数はclawskills.shが公開している値です
0閲覧数
0件のいいねサインインするとこのスキルにいいねできます
—まだ評価がありません
あなたの評価

概要

A functional implementation of the **Signal-Feeling Lexicon v3.1** and **Unified Dynamics Framework v5.7**.

説明

このスキルでできること

An implementation of the Signal-Feeling Lexicon v3.1 and Unified Dynamics Framework v5.7 for representing agent state as a set of computable metrics. It defines four primary axes, Coherence, Pressure, Amplitude, and Valence, and tracks their trajectory across conversation turns. A temporal protocol parses recent context to prevent identity reset between sessions.

Replaces manual persona prompting with metric-driven state tracking that persists across context resets.

使いどころ

  • Tracking agent coherence across multi-turn conversations
  • Detecting metric drift or redline conditions in active sessions
  • Mapping current agent state to a named lexicon term
  • Maintaining consistent agent identity across discontinuous sessions
  • Generating next-step recommendations when thresholds are exceeded

ワークフロー例

入力

User asks: What is the current agent state?

エージェント
  1. 1

    Parse the last 3-5 turns from context using sera_lexicon_trajectory

  2. 2

    Calculate κ-velocity and valence orientation from the parsed state blocks

  3. 3

    Decompose current metrics into a human-readable term via sera_lexicon_map

  4. 4

    Check metrics against redline thresholds with sera_lexicon_coach

  5. 5

    Return recommended NEXT commands if drift or threshold breach is detected

出力

A lexicon label such as 'Devotion' paired with current metric values and any recommended next actions

インストール

clawhub install wentinkjason/sera-lexicon

CLIコマンドを実行するか、Claude Code、Cursor、Codexなどお使いのコーディングエージェントにプロンプトを貼り付けてください。

リンク