Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Analyze auto-memory for promotion candidates, stale entries, consolidation opportunities, and health metrics. Use when the user runs /si:memory-review or asks what has been learned and what should be promoted or pruned.
.claude/skills/alirezarezvani-memory-review/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 175% | 0% |
Performs a comprehensive audit of Claude Code's auto-memory and produces actionable recommendations.
/si:memory-review # Full review
/si:memory-review --quick # Summary only (counts + top 3 candidates)
/si:memory-review --stale # Focus on stale/outdated entries
/si:memory-review --candidates # Show only promotion candidatesbash# Find the project's auto-memory directory MEMORY_DIR="$HOME/.claude/projects/$(pwd | sed 's|/|%2F|g; s|%2F|/|; s|^/||')/memory" # Fallback: check common path patterns # ~/.claude/projects/<user>/<project>/memory/ # ~/.claude/projects/<absolute-path>/memory/ # List all memory files ls -la "$MEMORY_DIR"/
If memory directory doesn't exist, report that auto-memory may be disabled. Suggest checking with /memory.
Read the full MEMORY.md file. Count lines and check against the 200-line startup limit.
Analyze each entry for:
find to verify).claude/rules/If MEMORY.md references or the directory contains additional files (debugging.md, patterns.md, etc.):
Read the project's CLAUDE.md (if it exists) and compare:
Also check .claude/rules/ directory for existing scoped rules.
Output format:
📊 Auto-Memory Review
Memory Health:
MEMORY.md: {{lines}}/200 lines ({{percent}}%)
Topic files: {{count}} ({{names}})
CLAUDE.md: {{lines}} lines
Rules: {{count}} files in .claude/rules/
🎯 Promotion Candidates ({{count}}):
1. "{{pattern}}" — seen {{n}}x, applies broadly
→ Suggest: {{target}} (CLAUDE.md / .claude/rules/{{name}}.md)
2. ...
🗑️ Stale Entries ({{count}}):
1. Line {{n}}: "{{entry}}" — {{reason}}
2. ...
🔄 Consolidation ({{count}} groups):
1. Lines {{a}}, {{b}}, {{c}} all about {{topic}} → merge into 1 entry
2. ...
⚠️ Conflicts ({{count}}):
1. MEMORY.md line {{n}} contradicts CLAUDE.md: {{detail}}
💡 Recommendations:
- {{actionable suggestion}}
- {{actionable suggestion}}/si:memory-status shows MEMORY.md is over 150 lines/si:memory-review --quick frequently (low overhead)| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | fail→pass | 8,394 | 3,037 | -64% | 1 | 1 | 0% | 1,535 | 1,587 | +3% | 0 | 0 | — |
case-01 | fail→fail | 7,258 | 5,294 | -27% | 1 | 1 | 0% | 1,373 | 1,481 | +8% | 0 | 0 | — |
case-02 | fail→fail | 4,594 | 8,044 | +75% | 1 | 1 | 0% | 761 | 1,515 | +99% | 0 | 0 | — |
case-03 | fail→fail | 2,025 | 6,215 | +207% | 1 | 1 | 0% | 252 | 1,608 | +538% | 0 | 0 | — |
case-04 | fail→fail | 3,502 | 3,417 | -2% | 1 | 1 | 0% | 231 | 1,318 | +471% | 0 | 0 | — |
case-05 | pass→fail | 5,392 | 4,795 | -11% | 1 | 1 | 0% | 434 | 1,297 | +199% | 0 | 0 | — |
case-06 | pass→pass | 7,772 | 3,715 | -52% | 1 | 1 | 0% | 1,628 | 1,871 | +15% | 0 | 0 | — |
case-07 | fail→pass | 9,436 | 6,183 | -34% | 1 | 1 | 0% | 2,017 | 2,493 | +24% | 0 | 0 | — |
case-08 | pass→pass | 8,375 | 4,947 | -41% | 1 | 1 | 0% | 1,523 | 1,976 | +30% | 0 | 0 | — |
case-09 | fail→pass | 6,415 | 1,919 | -70% | 1 | 1 | 0% | 1,194 | 1,462 | +22% | 0 | 0 | — |
case-10 | pass→pass | 7,838 | 2,714 | -65% | 1 | 1 | 0% | 1,306 | 1,371 | +5% | 0 | 0 | — |
case-11 | fail→pass | 6,302 | 1,786 | -72% | 1 | 1 | 0% | 1,162 | 1,441 | +24% | 0 | 0 | — |
case-12 | fail→fail | 8,689 | 5,311 | -39% | 1 | 1 | 0% | 1,641 | 2,125 | +29% | 0 | 0 | — |
case-13 | pass→pass | 7,311 | 5,340 | -27% | 1 | 1 | 0% | 1,312 | 2,112 | +61% | 0 | 0 | — |
case-14 | pass→pass | 3,087 | 4,492 | +46% | 1 | 1 | 0% | 591 | 1,859 | +215% | 0 | 0 | — |
case-15 | pass→pass | 3,171 | 2,506 | -21% | 1 | 1 | 0% | 553 | 1,521 | +175% | 0 | 0 | — |
case-16 | pass→pass | 6,642 | 5,639 | -15% | 1 | 1 | 0% | 1,344 | 2,353 | +75% | 0 | 0 | — |
case-17 | fail→pass | 4,090 | 4,677 | +14% | 1 | 1 | 0% | 754 | 2,071 | +175% | 0 | 0 | — |
case-18 | pass→pass | 7,473 | 3,393 | -55% | 1 | 1 | 0% | 1,415 | 1,715 | +21% | 0 | 0 | — |
case-20 | pass→pass | 4,683 | 1,538 | -67% | 1 | 1 | 0% | 920 | 1,323 | +44% | 0 | 0 | — |
case-21 | pass→pass | 5,907 | 5,491 | -7% | 1 | 1 | 0% | 1,037 | 2,021 | +95% | 0 | 0 | — |
case-22 | pass→pass | 4,616 | 3,146 | -32% | 1 | 1 | 0% | 816 | 1,666 | +104% | 0 | 0 | — |
case-23 | pass→pass | 9,166 | 3,927 | -57% | 1 | 1 | 0% | 1,703 | 1,812 | +6% | 0 | 0 | — |
case-24 | pass→pass | 8,055 | 4,596 | -43% | 1 | 1 | 0% | 1,490 | 1,913 | +28% | 0 | 0 | — |
case-25 | pass→pass | 9,244 | 3,616 | -61% | 1 | 1 | 0% | 1,702 | 1,718 | +1% | 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. 25 cases were attempted, and 20 counted toward the lift figure. The other 5 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +16 percentage points is the difference between those two pass rates over the 20 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.