Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Search Argentine legal databases (SAIJ, JUBA, CSJN, JUSCABA) for jurisprudence, legislation, case summaries, and doctrine using the `ley` CLI. Use when the user asks about Argentine law, court decisions, legal precedents, fallos, jurisprudencia, legislación, or mentions SAIJ, JUBA, CSJN, JUSCABA. Supports parallel search across databases, JSON/table/text output, and filtering by jurisdiction. Built from reverse-engineered MCP servers (hernan-cc) — direct HTTP calls, no MCP layer needed.
.claude/skills/thomasmoreai-ley-ar/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -53% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -55% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -51% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -52% | 0% |
Search jurisprudence, legislation, and doctrine from 4 public Argentine legal databases.
bashcd {baseDir}/scripts && pip install -e . --break-system-packages -q
Requires Python 3.10+. Dependencies: typer, httpx, rich.
| DB | Source | Best For | Reliability | |---|---|---|---| | saij | saij.gob.ar | National jurisprudence, legislation, doctrine | ✅ Clean JSON API | | csjn | sjconsulta.csjn.gov.ar | Supreme Court summaries | ✅ HTML+JSON | | juba | juba.scba.gov.ar | Buenos Aires Province decisions | ⚠️ HTML scraping | | juscaba | eje.juscaba.gob.ar | CABA court cases/expedientes | ⚠️ Poor free-text |
Default strategy: Start with --db saij,csjn. Add juba only for PBA-specific queries. Use juscaba only with case IDs.
bash# Search all databases (parallel) ley search "prescripción adquisitiva" # Filter by database(s) ley search "daño moral" --db saij,csjn # Limit results ley search "phishing bancario" --db saij --limit 5 # JSON output for scripting ley search "responsabilidad civil" --db csjn --json # Plain text ley search "contrato de locación" --text # Status check ley status ley --version
Each result: db, id, title, date, snippet, url. Default: Rich table. --json for structured data.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 15,901 | 7,080 | -55% | 1 | 1 | 0% | 2,913 | 832 | -71% | 0 | 0 | — |
case-02 | fail→fail | 22,511 | 7,147 | -68% | 1 | 1 | 0% | 3,423 | 761 | -78% | 0 | 0 | — |
case-03 | fail→fail | 22,174 | 7,961 | -64% | 1 | 1 | 0% | 3,449 | 818 | -76% | 0 | 0 | — |
case-04 | pass→pass | 12,753 | 63,088 | +395% | 1 | 1 | 0% | 2,394 | 1,147 | -52% | 0 | 0 | — |
case-05 | pass→pass | 8,691 | 2,061 | -76% | 1 | 1 | 0% | 1,478 | 869 | -41% | 0 | 0 | — |
case-06 | fail→pass | 15,016 | 6,661 | -56% | 1 | 1 | 0% | 2,728 | 1,343 | -51% | 0 | 0 | — |
case-07 | fail→pass | 11,617 | 1,962 | -83% | 1 | 1 | 0% | 1,885 | 882 | -53% | 0 | 0 | — |
case-08 | fail→pass | 15,906 | 3,441 | -78% | 1 | 1 | 0% | 2,717 | 1,211 | -55% | 0 | 0 | — |
case-09 | fail→pass | 8,945 | 1,294 | -86% | 1 | 1 | 0% | 1,511 | 736 | -51% | 0 | 0 | — |
case-10 | pass→pass | 10,484 | 2,936 | -72% | 1 | 1 | 0% | 1,736 | 1,022 | -41% | 0 | 0 | — |
case-11 | fail→pass | 9,859 | 2,139 | -78% | 1 | 1 | 0% | 1,861 | 900 | -52% | 0 | 0 | — |
case-12 | pass→pass | 13,292 | 12,656 | -5% | 1 | 1 | 0% | 2,163 | 2,684 | +24% | 0 | 0 | — |
case-13 | pass→pass | 9,782 | 7,185 | -27% | 1 | 1 | 0% | 1,942 | 1,639 | -16% | 0 | 0 | — |
case-14 | pass→pass | 9,380 | 4,037 | -57% | 1 | 1 | 0% | 1,719 | 873 | -49% | 0 | 0 | — |
case-15 | pass→pass | 7,123 | 1,205 | -83% | 1 | 1 | 0% | 1,197 | 673 | -44% | 0 | 0 | — |
case-16 | pass→pass | 8,176 | 1,652 | -80% | 1 | 1 | 0% | 1,398 | 768 | -45% | 0 | 0 | — |
case-17 | pass→pass | 10,066 | 1,884 | -81% | 1 | 1 | 0% | 1,741 | 832 | -52% | 0 | 0 | — |
case-18 | pass→pass | 9,846 | 1,483 | -85% | 1 | 1 | 0% | 1,621 | 713 | -56% | 0 | 0 | — |
case-19 | fail→pass | 14,234 | 5,679 | -60% | 1 | 1 | 0% | 2,217 | 1,438 | -35% | 0 | 0 | — |
case-20 | fail→fail | 9,854 | 10,716 | +9% | 1 | 1 | 0% | 1,714 | 2,405 | +40% | 0 | 0 | — |
case-21 | fail→fail | 6,998 | 10,019 | +43% | 1 | 1 | 0% | 617 | 1,535 | +149% | 0 | 0 | — |
case-22 | fail→pass | 8,702 | 3,220 | -63% | 1 | 1 | 0% | 1,498 | 1,081 | -28% | 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, and 19 counted toward the lift figure. The other 3 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 +32 percentage points is the difference between those two pass rates over the 19 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.