skills.rest
bluegitter
bluegitterCOMMUNITY
Explore all skills by this author →
GitHub

proactive-agent

Preserve session-critical details in SESSION-STATE.md with write-ahead logging and compaction recovery.

1|2|Updated Mar 21, 2026
Search Tags:#automation#security#agent#memory#proactive#self-healing#wal
One-click install
npx skills add https://github.com/bluegitter/QClaw --skill proactive-agent-bluegitter
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: proactive-agent
Source: https://github.com/bluegitter/QClaw/tree/main/resources/openclaw/config/skills/proactive-agent
Command: npx skills add https://github.com/bluegitter/QClaw --skill proactive-agent-bluegitter

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Proactive Agent addresses the problem of stateless, reactive conversational agents that lose context, miss opportunities, and fail silently. It provides patterns and file-based protocols so agents can anticipate user needs, survive context compaction, recover from truncated sessions, and operate with strong security guardrails.

Core Features & Use Cases

  • Write-Ahead Logging (WAL): Capture corrections, decisions, and important details to SESSION-STATE.md before responding.
  • Working Buffer & Compaction Recovery: Log danger-zone exchanges to memory/working-buffer.md and recover when context is truncated.
  • Autonomous Crons & Agent Orchestration: Distinguish systemEvent prompts from isolated agentTurns for reliable background tasks.
  • Security Hardening: Vet skill installation, scan for prompt injections, and run periodic security audits with provided scripts.
  • Self-Healing & Verification: Diagnose and fix issues, verify implementations (not just intent), and enforce anti-drift guardrails (ADL/VFM).
  • Use Case: Deploy for conversational AI that must maintain multi-session continuity, automatically audit security posture, and run autonomous maintenance without main-session interruptions.

Quick Start

Ask the proactive agent to recover the last session, run the security audit, and resume the top-priority task.

Frequently Asked Questions about proactive-agent

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

FAQPage Schema
How do I preserve conversational agent context across sessions when memory gets truncated?▼

To preserve conversational agent context, use write-ahead logging to capture critical details to a session state file before responding, and log exchanges to a working buffer for compaction recovery. This maintains state across truncated memory and sessions.

What is write-ahead logging for self-healing AI agents?▼

Write-ahead logging for self-healing agents is a protocol that records corrections and decisions to a file-based session state before the agent responds. This mechanism enables autonomous recovery and continuity during long or interrupted conversational workflows.

How do I set up autonomous cron tasks for background agent maintenance?▼

Set up autonomous cron tasks by distinguishing systemEvent prompts from isolated agentTurns. This separation allows reliable background maintenance and orchestration without interrupting the main conversational session.

Can I run a security audit on installed skills and scan for prompt injections?▼

Yes, you can run a security audit using provided periodic scripts to verify system integrity. The framework supports vetting skill installations and scanning for prompt injections to maintain a hardened security posture.

Does proactive agent recovery require a specific file format for session state?▼

Proactive agent recovery requires a file-based session state format, specifically a SESSION-STATE.md file, along with working-buffer logs. These files store the semantic memory needed to diagnose, self-heal, and resume tasks.

What is the best way to diagnose and verify AI agent implementations against drift?▼

The best way to verify implementations and prevent drift is to enforce anti-drift guardrails using verification mechanisms. Agents diagnose issues, fix them autonomously, and verify the actual implementation rather than just the original intent.

Related Skills

Explore More Software Engineering Skills→
zhengxuyuzhengxuyu

proactive-agent

Enable AI agents to anticipate needs and recover context using WAL and Working Buffer.

Community
Advanced
KrishnatejavepaKrishnatejavepa
14

doctor

Validate memory consistency, loop integrity, and secret safety in a cofounder system.

Community
Advanced
NinatuziNinatuzi

cron-scheduler

Schedule cron jobs with collision prevention and quiet hours gating.

Community
Advanced
Wike-CHIWike-CHI
2

proactive-agent

Persist session-critical decisions and context to SESSION-STATE.md before responding.

Community
Advanced