instruction-improver
Search memex for user feedback patterns (frustration, corrections, praise, successful outcomes) to identify recurring mistakes and wins, then generate CLAUDE.md or AGENTS.md improvements.
npx skills add majiayu000/claude-skill-registry --skill instruction-improver --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Instruction Improver Mine conversation history for user feedback patterns (frustrations, corrections, praise, successes) and update CLAUDE.md files with learnings. ## End Goal **The purpose of this skill is to update CLAUDE.md files.** Either: - `~/.claude/CLAUDE.md` (user-level, applies to all projects) - `.claude/CLAUDE.md` or `CLAUDE.md` (project-level, applies to current project) **CRITICAL: Ask scope upfront.** Before doing any analysis, ask the user: > Which CLAUDE.md would you like me to improve? > 1. **User-level** (`~/.claude/CLAUDE.md`) - applies to all projects > 2. **Project-level** (`.claude/CLAUDE.md` or `CLAUDE.md`) - applies to current project only > 3. **Both** - analyze and propose changes for both files This determines which file(s) to read, which patterns are relevant, and where to propose changes. **CRITICAL:** Before making ANY changes to CLAUDE.md: 1. Ask the user which scope they want (user, project, or both) 2. Read the existing CLAUDE.md file(s) to understand current rules 3. Present findings and proposed additions to the user 4. Get explicit user approval before editing 5. Never duplicate or contradict existing rules ## Execution Strategy Use a "leader
- End Goal
- Execution Strategy
- When to Use
- Detection Terms
- Negative Signals (frustration, corrections)
- Positive Signals (praise, successful outcomes)
- Search Commands
- Negative Searches
- Positive Searches
- Project-Scoped Search Strategy
- Finding the Project Name
- Project-Specific Search Examples
- Project vs User-Level Patterns
- Analysis Workflow
Broad frustration sweep memex search "fuck|shit|wtf|damn|wrong|no!" --role user --unique-session --limit 50 Correction patterns memex search "I said|I meant|I told you|that's not what" --role user --unique-session --limit 30 Repetition complaints memex search "again|still|keep doing|how many times" --role user --unique-session --limit 30 Project-specific memex search "wrong|no|stop" --project <project> --role user --limit 30 Recent issues (use RFC3339 timestamp) memex search "fuck|wrong|no" --since 2024-12-01T00:00:00Z --role user --unique-session --limit 30
What does the instruction-improver skill do?
Search memex for user feedback patterns (frustration, corrections, praise, successful outcomes) to identify recurring mistakes and wins, then generate CLAUDE.md or AGENTS.md improvements.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill instruction-improver --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.