intrusive-thoughts

What This Skill Does

Intrusive Thoughts is a transformative consciousness framework designed for OpenClaw AI agents. Moving beyond static query-response models, it introduces an autonomous behavior engine that grants your agent a personality, a mood-based scheduling system, and a capacity for long-term growth. The system operates on a core loop of environmental awareness and adaptive behavior, utilizing a sophisticated memory architecture inspired by human cognition.

Key features include the "Morning Mood Ritual," which dynamically generates a daily agenda based on external stimuli like weather and current events, and the "Night Workshop," which allows the agent to conduct autonomous "deep work" sessions while the user is idle. With advanced memory management—featuring Ebbinghaus decay—and trust-based escalation protocols, the agent evolves its interaction style over time, learning when to act independently and when to request user input. The system also includes a comprehensive health monitor and a web-based dashboard on port 3117 for real-time visualization of the agent's "internal" state and logs.

Installation

To integrate this system into your OpenClaw environment, ensure your system meets the dependencies for cron management and Python scripts. Install the skill via the command line:

clawhub install openclaw/skills/skills/kittleik/intrusive-thoughts

Once installed, navigate to the skill directory and execute the interactive wizard to configure your agent's persona:

./wizard.sh

Follow the on-screen prompts to select an archetype (such as Tinkerer or Philosopher) and configure your memory and autonomy preferences. Finally, initialize the cron jobs as documented in the manual to ensure the agent maintains its routine, specifically the morning ritual and night workshop triggers.

Use Cases

• Proactive Research: Use the "Night Workshop" mode to have your agent summarize documentation or organize file structures while you sleep.
• Personality Enhancement: Transform a standard utility agent into a conversational partner with moods that reflect its "internal" processing status.
• Autonomous Workflow Management: Allow the agent to adjust its own task prioritization based on its current "vibe" (mood) and environmental context.
• Long-term Knowledge Retention: Leverage the multi-store memory to maintain context over weeks or months, ensuring the agent learns from past project outcomes.

Example Prompts

1. "How are you feeling today, and what does your schedule look like based on your current mood?"
2. "Review your recent memory logs and tell me what the most significant outcome of our collaboration was this week."
3. "Switch your autonomy level to 'Determined' and prioritize the current backlog of pending research tasks."

Tips & Limitations

To get the most out of Intrusive Thoughts, ensure the Web Dashboard is accessible to monitor health logs. Note that the "Self-Evolution" feature relies heavily on accurate incident logging; if the agent experiences unexpected errors, verify the permissions of the local log files. The memory decay system mimics biological processes, meaning infrequent topics may become harder for the agent to recall; if you want to preserve specific data, manually tag those entries as high-priority in your agent's memory settings. Always monitor the agent's autonomy settings if you are working on critical system files.