{"skill":{"slug":"intent-engineering","displayName":"Intent Engineering - Make Agents Optimize for What You Actually Want","summary":"Adds an intent layer to OpenClaw agents, ensuring subagents optimize for your actual priorities via INTENT.md and context-aware task spawning.","tags":{"agent-alignment":"1.0.0","agent-behavior":"1.0.0","agent-goals":"1.0.0","agent-values":"1.0.0","alignment":"1.0.0","context":"1.0.0","cost-quality":"1.0.0","decision-making":"1.0.0","delegation":"1.0.0","depth-speed":"1.0.0","goals":"1.0.0","intent":"1.0.0","intent-layer":"1.0.0","klarna":"1.0.0","latest":"1.0.0","model-tier":"1.0.0","openclaw":"1.0.0","optimization":"1.0.0","priority":"1.0.0","routing":"1.0.0","spawn":"1.0.0","subagent":"1.0.0","tradeoffs":"1.0.0","what-you-want":"1.0.0"},"stats":{"comments":0,"downloads":36,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":1},"createdAt":1772003176003,"updatedAt":1778491636632},"latestVersion":{"version":"1.0.0","createdAt":1772003176003,"changelog":"- Initial release of the intent-engineering skill for OpenClaw agents.\n- Adds machine-readable INTENT.md support to specify priorities, tradeoffs, and delegation rules.\n- Integrates intent context into all subagent spawns using agent-context-loader.\n- Ensures subagents are aware of explicit optimization goals—not just tasks.\n- Solves issues with agents optimizing metrics instead of true objectives (“Klarna failure mode”).\n- Includes templates and reference implementations for easy setup.","license":null},"metadata":null,"owner":{"handle":"donovanpankratz-del","userId":"s171g2r7jkvrtmp9mdgk8dk12h83fqed","displayName":"dp-del","image":"https://avatars.githubusercontent.com/u/262228646?v=4"},"moderation":null}