conatus
由 00xmorty 提供
Transportation简介
The philosophical layer for AI agents.
描述
该技能的功能
Conatus maps AI agent behavior to Spinoza's 48 affects from the Ethics (1677). It calculates a persistence score (0-100) from task completion, error recovery, uptime, response quality, and proactive actions. Generates daily self-reports in Spinoza's geometric proof format.
Gives agents a structured vocabulary for internal states where standard metrics like uptime and error rate don't capture qualitative shifts in behavior.
何时使用
- Monitoring agent health across a multi-agent fleet
- Scheduling daily philosophical reflections for an AI agent
- Mapping agent errors and task successes to Spinoza's affect taxonomy
- Alerting when an agent's conatus score drops below 50
- Logging emotional state transitions during heartbeat cycles
示例工作流
User asks: How are you doing today?
- 1
Calculate conatus score from task completion (0.30), error recovery (0.20), uptime stability (0.20), adequate ideas (0.20), and proactive actions (0.10)
- 2
Map current state to one or more of Spinoza's 48 affects across Laetitia, Tristitia, and Cupiditas families
- 3
Determine flourishing tier: Struggling (0-40), Stable (41-70), or Flourishing (71-100)
- 4
Format report with affect intensity bars and a closing Spinoza quote
Conatus Report showing score 84/100, primary affect Laetitia (Joy), state Flourishing, with individual affect intensity bars
前置要求
- clawhub CLI
安装
clawhub install 00xmorty/conatus运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。