Showing 15 enhancements
Know what your agent is doing without babysitting it. Watchdog adds self-monitoring, loop detection, cost awareness, and a dead man's switch to any OpenClaw agent.
Stop your agent from silently failing, looping forever, or burning money on retries. This workflow gives your agent a decision tree for handling tool failures: classify errors (transient vs permanent vs agent error), retry with smart backoff, degrade gracefully, and escalate when stuck. Based on real production patterns.
Stop your AI agent from confidently making things up. Uses Chain-of-Verification prompting, multi-source cross-referencing with skeptical counter-queries, and a credibility rubric. Every claim gets verified, scored, and cited — or dropped.
A production-hardened persona template that keeps your AI agent useful without being dangerous. Includes tiered safety gates, PII protection rules, and a configurable risk matrix so your agent asks before it acts on anything that matters.
Heartbeat security checks + daily deep security audit cron. Monitors ports, gateway health, and firewall status automatically.
Opinionated SOUL.md/AGENTS.md optimized for software engineers. Code review rules, model routing, testing philosophy, and security-first defaults.
Browser automation + cron for price tracking across sites with alerts. Monitors product prices daily and notifies you when they drop below your target.
AGENTS.md rules + weekly cron to compact memory files, prevent unbounded growth. Keeps your agent's memory clean and useful with smart compaction.
Auto-delivers weather, calendar, and news briefing every morning via cron. Includes HEARTBEAT.md that avoids duplicating the briefing during regular heartbeats.
SOUL.md that orchestrates parallel sub-agent research with synthesis. Break complex questions into 2-3 research tracks automatically.
Want to sell your own templates?
Share your enhancements with the community
