From method-hoard
Actively reflect on recent work to discover generalizable methods worth hoarding
How this skill is triggered — by the user, by Claude, or both
Slash command
/method-hoard:discoverThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use the Agent tool to launch the `method-hoard:discover` agent with model `opus`. Pass `$ARGUMENTS` as context for what to reflect on (file paths, topic, "diff", or empty for broad reflection).
Use the Agent tool to launch the method-hoard:discover agent with model opus. Pass $ARGUMENTS as context for what to reflect on (file paths, topic, "diff", or empty for broad reflection).
When the agent returns with candidates, present each candidate to the user via AskUserQuestion. For each candidate, offer three options:
/method-hoard:stock with the candidate detailsThen act on the user's choices: invoke /method-hoard:stock for each "Stock" selection, discuss any "Discuss" selections, and move on past any "Skip" selections.
npx claudepluginhub davidorex/method-hoardExplores codebases, researches best practices, evaluates 2-3 approaches with pros/cons, and recommends solutions with codebase-specific reasoning before planning.
Reflects on recent work, scans SHELL.md for patterns, and proposes improvements. Useful for periodic code quality and process reflection.
Analyzes conversation history to propose additions to project's CLAUDE.md, capturing code conventions, architecture decisions, tool preferences, and project-specific rules. Use after sessions to document learnings.