Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Execute autonomous multi-step research using Google Gemini Deep Research Agent. Use for: market analysis, competitive landscaping, literature reviews, technical research, due diligence. Takes 2-10 minutes but produces detailed, cited reports. Costs $2-5 per task.
.claude/skills/sanjay3290-deep-research/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | -62% | 0% |
| case-19 | ✗→✓ | ▲ Improved | -49% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -25% | 0% |
| case-16 | ✗→✓ | ▲ Improved | -65% | 0% |
Run autonomous research tasks that plan, search, read, and synthesize information into comprehensive reports.
pip install -r requirements.txtbash export GEMINI_API_KEY=your-api-key-here Or create a .env file in the skill directory.
bashpython3 scripts/research.py --query "Research the history of Kubernetes"
bashpython3 scripts/research.py --query "Compare Python web frameworks" \ --format "1. Executive Summary\n2. Comparison Table\n3. Recommendations"
bashpython3 scripts/research.py --query "Analyze EV battery market" --stream
bashpython3 scripts/research.py --query "Research topic" --no-wait
bashpython3 scripts/research.py --status <interaction_id>
bashpython3 scripts/research.py --wait <interaction_id>
bashpython3 scripts/research.py --query "Elaborate on point 2" --continue <interaction_id>
bashpython3 scripts/research.py --list
--json): Structured data for programmatic use--raw): Unprocessed API response| Metric | Value | |--------|-------| | Time | 2-10 minutes per task | | Cost | $2-5 per task (varies by complexity) | | Token usage | ~250k-900k input, ~60k-80k output |
--query "..."--stream or poll with --status--continue for follow-up questions| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | fail→fail | 5,458 | 4,663 | -15% | 1 | 1 | 0% | 860 | 922 | +7% | 0 | 0 | — |
case-05 | fail→fail | 4,963 | 5,188 | +5% | 1 | 1 | 0% | 726 | 986 | +36% | 0 | 0 | — |
case-01 | fail→fail | 27,988 | 5,595 | -80% | 1 | 1 | 0% | 4,825 | 962 | -80% | 0 | 0 | — |
case-02 | fail→pass | 15,152 | 6,881 | -55% | 1 | 1 | 0% | 2,644 | 1,005 | -62% | 0 | 0 | — |
case-03 | fail→fail | 27,875 | 8,306 | -70% | 1 | 1 | 0% | 5,519 | 915 | -83% | 0 | 0 | — |
case-04 | fail→fail | 19,913 | 5,008 | -75% | 1 | 1 | 0% | 3,288 | 964 | -71% | 0 | 0 | — |
case-07 | fail→fail | 2,505 | 4,166 | +66% | 1 | 1 | 0% | 360 | 920 | +156% | 0 | 0 | — |
case-08 | fail→fail | 2,102 | 7,637 | +263% | 1 | 1 | 0% | 308 | 1,753 | +469% | 0 | 0 | — |
case-09 | fail→fail | 19,200 | 5,226 | -73% | 1 | 1 | 0% | 2,284 | 983 | -57% | 0 | 0 | — |
case-10 | pass→pass | 8,889 | 2,156 | -76% | 1 | 1 | 0% | 1,634 | 1,020 | -38% | 0 | 0 | — |
case-19 | fail→pass | 10,464 | 1,802 | -83% | 1 | 1 | 0% | 1,813 | 928 | -49% | 0 | 0 | — |
case-11 | fail→pass | 11,489 | 3,445 | -70% | 1 | 1 | 0% | 1,875 | 1,184 | -37% | 0 | 0 | — |
case-12 | fail→pass | 6,305 | 948 | -85% | 1 | 1 | 0% | 1,040 | 781 | -25% | 0 | 0 | — |
case-13 | pass→pass | 4,144 | 1,329 | -68% | 1 | 1 | 0% | 769 | 851 | +11% | 0 | 0 | — |
case-14 | pass→pass | 8,501 | 1,365 | -84% | 1 | 1 | 0% | 1,332 | 800 | -40% | 0 | 0 | — |
case-15 | pass→pass | 3,619 | 1,072 | -70% | 1 | 1 | 0% | 559 | 777 | +39% | 0 | 0 | — |
case-16 | fail→pass | 12,378 | 1,079 | -91% | 1 | 1 | 0% | 2,308 | 813 | -65% | 0 | 0 | — |
case-17 | fail→pass | 7,135 | 1,360 | -81% | 1 | 1 | 0% | 1,221 | 780 | -36% | 0 | 0 | — |
case-18 | pass→pass | 14,768 | 1,503 | -90% | 1 | 1 | 0% | 2,601 | 895 | -66% | 0 | 0 | — |
case-20 | fail→fail | 3,234 | 10,675 | +230% | 1 | 1 | 0% | 501 | 2,359 | +371% | 0 | 0 | — |
case-21 | pass→pass | 1,818 | 2,192 | +21% | 1 | 1 | 0% | 300 | 985 | +228% | 0 | 0 | — |
case-22 | pass→pass | 2,547 | 2,235 | -12% | 1 | 1 | 0% | 402 | 916 | +128% | 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 16 counted toward the lift figure. The other 6 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 +27 percentage points is the difference between those two pass rates over the 16 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.