borghei/Claude-Skills / engineering/self-improving-agent
self-improving-agent
VerifiedCurates agent memory and turns repeated lessons into reviewed rules: memory health checks, pattern extraction, feedback and regression analysis. Use when managing auto-memory or building agents that learn from their own sessions.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.0 KB |
| assets/memory_index/MEMORY.md | markdown | 316 B |
| assets/memory_index/feedback_migrations.md | markdown | 203 B |
| assets/memory_index/feedback_pnpm.md | markdown | 255 B |
| assets/sample_feedback.jsonl | other | 2.2 KB |
| assets/sample_metrics.json | config | 3.6 KB |
| references/feedback-loop-patterns.md | markdown | 9.2 KB |
| references/memory-curation-guide.md | markdown | 4.6 KB |
| references/meta-learning-architectures.md | markdown | 2.4 KB |
| references/self-improvement-methodology.md | markdown | 4.3 KB |
| scripts/feedback_analyzer.pyexecutable | script | 18.5 KB |
| scripts/memory_health_checker.py | script | 13.1 KB |
| scripts/pattern_extractor.py | script | 10.8 KB |
| scripts/regression_detector.pyexecutable | script | 17.0 KB |
| scripts/rule_manager.pyexecutable | script | 12.0 KB |
| scripts/rule_promoter.py | script | 12.0 KB |
| skills/extract.md | markdown | 2.1 KB |
| skills/promote.md | markdown | 2.8 KB |
| skills/remember.md | markdown | 2.1 KB |
| skills/review.md | markdown | 2.3 KB |
| skills/status.md | markdown | 2.5 KB |
skills/promote.md
View on GitHub# Sub-Skill: Promote
**Parent:** self-improving-agent
**Trigger:** "promote pattern to rule", "graduate to CLAUDE.md", "make this permanent"
## Purpose
Graduate proven patterns from memory (MEMORY.md) to enforced rules (CLAUDE.md or `.claude/rules/`). This is the critical step that turns observations into permanent behavior changes.
## Workflow
### Step 1: Check Promotion Criteria
A pattern is ready for promotion when ALL of these are met:
| Criterion | Threshold | Verification |
|-----------|-----------|--------------|
| Recurrence | 3+ sessions | Check recurrence count in memory |
| Consistency | Same solution every time | No contradicting entries exist |
| Impact | Prevented errors or saved time | At least one error prevented |
| Stability | Underlying system unchanged | Referenced code/tools still exist |
| Clarity | Statable in 1-2 sentences | Can be expressed as a clear rule |
### Step 2: Determine Target
| Pattern Type | Promote To | Format |
|-------------|-----------|--------|
| Coding convention | `.claude/rules/<area>.md` | Rule with scope path |
| Project architecture | `CLAUDE.md` | Architecture section entry |
| Tool preference | `CLAUDE.md` | Development environment section |
| Debugging pattern | `.claude/rules/debugging.md` | Conditional rule |
| File-scoped rule | `.claude/rules/<scope>.md` with `paths:` | Scoped rule |
### Step 3: Draft the Rule
Format as a clear, enforceable statement:
- Start with an action verb (Use, Always, Never, Prefer)
- Include the "why" as a brief annotation
- Scope to specific files/directories if applicable
Example:
```
Always use `type` not `interface` for object shapes.
Reason: Consistent with codebase convention; types are more flexible for unions.
```
### Step 4: Apply Promotion
Preview with `rule_promoter.py`, show the user the rule text and target, and apply only after an explicit yes (a promoted rule shapes every later session):
```bash
python3 scripts/rule_promoter.py --memory ./MEMORY.md --entry-title "Use pnpm" --target claude-md --target-path ./CLAUDE.md --dry-run
python3 scripts/rule_promoter.py --memory ./MEMORY.md --entry-title "Use pnpm" --target claude-md --target-path ./CLAUDE.md --apply
```
### Step 5: Clean Up Memory
After promotion:
- Remove the original entry from MEMORY.md
- Add a reference note: "Promoted to CLAUDE.md on [date]"
- Verify MEMORY.md line count is within limits
## Inputs
| Input | Required | Description |
|-------|----------|-------------|
| Entry title | Yes | Title (or part of it) of the `## ` memory entry to promote |
| Target | Yes | CLAUDE.md or .claude/rules/<name>.md |
| Dry run | No | Preview without applying (default: true) |
## Outputs
- Promotion validation result (pass/fail with reasons)
- Draft rule text
- Applied changes (if not dry run)
- Updated memory with entry removed
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |