From agent-memory
Create a lesson — a proactive rule that fires BEFORE risky operations. Unlike observations (passive), lessons are instructions injected at session start and triggered by PreToolUse hooks. Use after learning from a mistake.
How this skill is triggered — by the user, by Claude, or both
Slash command
/agent-memory:create_lesson <rule> [--title VALUE] [--severity VALUE] [--project VALUE] [--trigger_tool VALUE] [--trigger_pattern VALUE]<rule> [--title VALUE] [--severity VALUE] [--project VALUE] [--trigger_tool VALUE] [--trigger_pattern VALUE]The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Create a lesson — a proactive rule that fires BEFORE risky operations. Unlike observations (passive), lessons are instructions injected at session start and triggered by PreToolUse hooks. Use after learning from a mistake.
Create a lesson — a proactive rule that fires BEFORE risky operations. Unlike observations (passive), lessons are instructions injected at session start and triggered by PreToolUse hooks. Use after learning from a mistake.
| Param | Type | Required | Description |
|---|---|---|---|
rule | string | Yes | The instruction/rule (e.g. 'ALWAYS diff dev vs prod config before deploying') |
title | string | No | Short title for the lesson |
severity | string | No | How important (default: warning) |
project | string | Yes | Project path (full cwd) — ALWAYS pass the current project path. Omit ONLY for truly global lessons. |
trigger_tool | string | No | Tool to match: Bash, Edit, Write, NotebookEdit (omit for any) |
trigger_pattern | string | No | Regex to match against tool input (e.g. 'amplify.*update-app') |
When invoked with /create_lesson, call the create_lesson MCP tool (server: agent-memory) with rule="$ARGUMENTS".
IMPORTANT: Always include the project parameter set to the current project path (full cwd from the session-start context).
create_lesson(rule="$ARGUMENTS", project="<current_project>")
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
npx claudepluginhub metazen11/agent-memory --plugin agent-memory