agent-memory-discipline - Standing rules for agent memory recall and save
Establishes standing rules for when an agent reads long-term memory before acting and writes decisions, corrections, and failures afterwards.
Tags
Updated: 2026-09-24Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Recall memory before taking action
- Save decisions, corrections, and failures
- Close superseded memory entries with dates
- Surface conflicting memory entries
Inputs
- User prompts and context
- Connected memory backend or server
Outputs
- Recalled context with sources and dates
- Structured memory entries
Requirements
- Connected memory tool or MCP backend