Stats
Actions
Tags
From lexic
Record a learning that will be served as context in future autonomous runs
How this command is triggered — by the user, by Claude, or both
Slash command
/lexic:learn what you learned and why it mattersThe summary Claude sees in its command listing — used to decide when to auto-load this command
# Record Learning Capture a specific, actionable learning that should influence future autonomous runs. This is different from `/Lexic:save` — saved knowledge is for general reference, while learnings are actively assembled into the context package when runs work on related features. Use this when you've figured something out **outside** of a run and want future runs to benefit from it. ## Steps 1. Take `$ARGUMENTS` as the raw learning. 2. Assess the quality against these criteria: - **Specific**: Does it name concrete things (files, functions, error messages, versions)? - **Actio...
Capture a specific, actionable learning that should influence future autonomous runs. This is different from /Lexic:save — saved knowledge is for general reference, while learnings are actively assembled into the context package when runs work on related features.
Use this when you've figured something out outside of a run and want future runs to benefit from it.
Take $ARGUMENTS as the raw learning.
Assess the quality against these criteria:
If the learning is vague (e.g., "auth is complicated" or "watch out for the database"), ask the user to add specifics. Prompt with:
Once the learning is specific enough, ask the user:
critical = must always surface, useful = surface when relevant)Call workflow_learning_create with:
content: The learning, formatted clearlyfeature_refs: Array of feature or component names this relates to (e.g. ["billing", "oauth"])category: The type of learning (e.g. gotcha, best_practice, security, general)context_tags: Array of tags for context matching (e.g. ["supabase", "rls"])confidence: 1.0 for critical learnings, 0.5 for useful onesConfirm what was recorded.
| Situation | Use |
|---|---|
| "Supabase RLS silently fails if you join auth.users" | /Lexic:learn — this should change how future runs write RLS policies |
| "Meeting notes from the API design review" | /Lexic:save — reference material, not a run-level insight |
| "The billing webhook needs idempotency keys" | /Lexic:learn — future runs touching billing should know this |
| "List of competitor features we analyzed" | /Lexic:save — research, not an implementation learning |
npx claudepluginhub lexic-io/lexic.io-plugin --plugin lexic/learnAnalyzes the current session for non-trivial patterns—error resolutions, debugging techniques, workarounds, and conventions—then saves them as reusable skill files.
/learnInteractively explores Task Master capabilities with tailored suggestions, command lists by category/scenario, power user patterns, and learning paths based on arguments or project state.
/learnAutonomously scouts codebase, learns structure, generates or updates documentation with validation-fix loops.
/learnAnalyzes the current session for non-trivial patterns—error resolutions, debugging techniques, workarounds, and conventions—then saves them as reusable skill files.
/learnDisplays Claude Code best practices guide on sessions, context, memory, modes, CLI shortcuts, worktrees, and prompting. Also supports specific topics and saving session lessons to persistent memory.
/learnToggles learning mode in vibe sessions to enable/disable educational micro-explanations after each step. Accepts 'on', 'off', or no argument to toggle.