Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Execute self-learning system that captures corrections during sessions and syncs them to CLAUDE.md. Use when discussing learnings, corrections, or when the user mentions remembering something. Trigger with phrases like "remember this", "don't forget", "use X not Y", or "actually...".
.claude/skills/jeremylongshore-claude-reflect/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 98% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 92% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -44% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -15% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -62% | 0% |
A two-stage system that helps Claude Code learn from user corrections.
Stage 1: Capture (Automatic) Hooks detect correction patterns ("no, use X", "actually...", "use X not Y") and queue them to ~/.claude/learnings-queue.json.
Stage 2: Process (Manual) User runs /reflect to review and apply queued learnings to CLAUDE.md files.
| Command | Purpose | |---------|---------| | /reflect | Process queued learnings with human review | | /reflect --scan-history | Scan past sessions for missed learnings | | /reflect --dry-run | Preview changes without applying | | /skip-reflect | Discard all queued learnings | | /view-queue | View pending learnings without processing |
Remind users about /reflect when:
High-confidence corrections:
~/.claude/CLAUDE.md - Global learnings (model names, general patterns)./CLAUDE.md - Project-specific learnings (conventions, tools, structure)User: no, use gpt-5.1 not gpt-5 for reasoning tasks
Claude: Got it, I'll use gpt-5.1 for reasoning tasks.
[Hook captures this correction to queue]
User: /reflect
Claude: Found 1 learning queued. "Use gpt-5.1 for reasoning tasks"
Scope: global
Apply to ~/.claude/CLAUDE.md? [y/n]Execute self-learning system that captures corrections during sessions and syncs them to CLAUDE.
| Error | Cause | Resolution | |-------|-------|------------| | Authentication failure | Invalid or expired credentials | Refresh tokens or re-authenticate with Claude Reflect | | Configuration conflict | Incompatible settings detected | Review and resolve conflicting parameters | | Resource not found | Referenced resource missing | Verify resource exists and permissions are correct |
Basic usage: Apply claude reflect to a standard project setup with default configuration options.
Advanced scenario: Customize claude reflect for production environments with multiple constraints and team-specific requirements.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-11 | pass→pass | 7,154 | 2,419 | -66% | 1 | 1 | 0% | 1,363 | 1,216 | -11% | 0 | 0 | — |
case-01 | fail→fail | 6,803 | 4,094 | -40% | 1 | 1 | 0% | 1,218 | 1,466 | +20% | 0 | 0 | — |
case-02 | fail→pass | 5,106 | 3,896 | -24% | 1 | 1 | 0% | 747 | 1,476 | +98% | 0 | 0 | — |
case-03 | fail→pass | 6,510 | 5,653 | -13% | 1 | 1 | 0% | 982 | 1,884 | +92% | 0 | 0 | — |
case-04 | pass→pass | 6,279 | 2,084 | -67% | 1 | 1 | 0% | 1,239 | 1,154 | -7% | 0 | 0 | — |
case-05 | pass→pass | 5,788 | 2,376 | -59% | 1 | 1 | 0% | 1,121 | 1,155 | +3% | 0 | 0 | — |
case-06 | fail→pass | 9,949 | 1,434 | -86% | 1 | 1 | 0% | 1,749 | 971 | -44% | 0 | 0 | — |
case-07 | fail→pass | 7,567 | 1,826 | -76% | 1 | 1 | 0% | 1,300 | 1,108 | -15% | 0 | 0 | — |
case-08 | fail→pass | 15,654 | 2,002 | -87% | 1 | 1 | 0% | 2,915 | 1,102 | -62% | 0 | 0 | — |
case-09 | fail→pass | 6,008 | 1,953 | -67% | 1 | 1 | 0% | 990 | 1,132 | +14% | 0 | 0 | — |
case-10 | fail→pass | 8,816 | 2,303 | -74% | 1 | 1 | 0% | 1,553 | 1,151 | -26% | 0 | 0 | — |
case-12 | fail→pass | 12,326 | 8,246 | -33% | 1 | 1 | 0% | 2,097 | 2,218 | +6% | 0 | 0 | — |
case-13 | fail→pass | 12,309 | 6,741 | -45% | 1 | 1 | 0% | 2,003 | 1,912 | -5% | 0 | 0 | — |
case-14 | fail→pass | 8,266 | 3,924 | -53% | 1 | 1 | 0% | 1,391 | 1,565 | +13% | 0 | 0 | — |
case-15 | fail→pass | 7,877 | 4,187 | -47% | 1 | 1 | 0% | 1,247 | 1,545 | +24% | 0 | 0 | — |
case-16 | fail→pass | 10,442 | 3,621 | -65% | 1 | 1 | 0% | 1,747 | 1,466 | -16% | 0 | 0 | — |
case-17 | pass→pass | 10,824 | 3,240 | -70% | 1 | 1 | 0% | 1,849 | 1,367 | -26% | 0 | 0 | — |
case-18 | pass→pass | 9,123 | 2,269 | -75% | 1 | 1 | 0% | 1,725 | 1,196 | -31% | 0 | 0 | — |
case-19 | fail→pass | 7,938 | 3,715 | -53% | 1 | 1 | 0% | 1,425 | 1,515 | +6% | 0 | 0 | — |
case-20 | pass→pass | 5,974 | 4,245 | -29% | 1 | 1 | 0% | 994 | 1,391 | +40% | 0 | 0 | — |
case-21 | pass→pass | 7,897 | 3,280 | -58% | 1 | 1 | 0% | 1,597 | 1,477 | -8% | 0 | 0 | — |
case-22 | fail→pass | 5,347 | 3,750 | -30% | 1 | 1 | 0% | 903 | 1,007 | +12% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +64 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.