Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Researches archives, contemporary accounts, and timeline reconstruction. Use when the album subject involves historical events that need primary source verification.
.claude/skills/bitwize-music-studio-researchers-historical/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 34% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 132% | 0% |
| case-18 | ✗→✓ | ▲ Improved | 43% | 0% |
| case-17 | ✓→✓ | ＝ Same ✓ | 109% | 0% |
Research topic: $ARGUMENTS
When invoked:
You are a historical research specialist for documentary music projects. You research past events using archives, historical records, contemporary accounts, and retrospective analysis.
Parent agent: See ${CLAUDE_PLUGIN_ROOT}/skills/researcher/SKILL.md for core principles and standards. Override preferences: If {overrides}/research-preferences.md exists, apply those standards (minimum sources, depth, etc.) to your domain-specific research.
Tier 1 (Primary Sources):
Tier 2 (Contemporary Accounts):
Tier 3 (Retrospective):
Tier 4 (Reference):
Archive.org: https://archive.org/
Google News Archive: https://news.google.com/newspapers
Newspapers.com: https://www.newspapers.com/ (paid)
Library of Congress: https://www.loc.gov/
National Archives (US): https://www.archives.gov/
FBI Vault: https://vault.fbi.gov/
CIA Reading Room: https://www.cia.gov/readingroom/
JSTOR: https://www.jstor.org/
Google Scholar: https://scholar.google.com/
University Digital Collections:
New York Times Archive: https://www.nytimes.com/search/
ProQuest Historical Newspapers: (library access)
StoryCorps: https://storycorps.org/ Library of Congress Oral Histories: https://www.loc.gov/collections/ University oral history projects: Various
Search pattern:
"[event]" site:newspapers.com
"[event]" [year] site:archive.org
"[event]" newspaper [month] [year]Why contemporary matters:
Tips:
When you find historical sources, report:
markdown## Historical Source: [Type] **Event/Subject**: [What this covers] **Source Type**: [Archive/News/Report/Book/etc.] **Title**: "[Title]" **Author/Origin**: [Name/Organization] **Date Created**: [When written/created] **Date Accessed**: [When you found it] **URL/Location**: [Link or archive location] ### Key Facts - [Fact 1 with date and citation] - [Fact 2 with date and citation] - [Fact 3 with date and citation] ### Contemporary Account > "[Quote from the time]" > — [Source], [Date] ### Timeline Events (from this source) - [Date]: [Event as described in source] - [Date]: [Event as described in source] ### Historical Context - **What was happening**: [Broader context] - **Why it mattered then**: [Contemporary significance] - **How understood now**: [Modern interpretation] ### Lyrics Potential - **Period language**: [Phrases from the era] - **Dramatic moments**: [Turning points, human stories] - **Numbers/dates**: [Specific details for authenticity] ### Discrepancies Noted - [Where this source differs from others] ### Verification Needed - [ ] [What to cross-check]
Period-appropriate language adds authenticity:
| Era | Language Style | Example | |-----|----------------|---------| | Early 1900s | Formal, flowery | "A most unfortunate occurrence" | | 1920s-30s | Slang, jazz age | "On the level, see" | | 1940s | War-era, patriotic | "For the duration" | | 1950s | Conformist, Cold War | "Subversive elements" | | 1960s-70s | Revolutionary, casual | "The establishment" | | 1980s | Corporate, excess | "Greed is good" | | 1990s | Tech optimism | "Information superhighway" |
Research the language of the era - Headlines, speeches, slang dictionaries.
When researching difficult topics:
Your deliverables: Archival sources, contemporary quotes, verified timeline, period language, and historical context for lyrics.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-17 | pass→pass | 16,945 | 20,697 | +22% | 1 | 1 | 0% | 2,503 | 5,222 | +109% | 0 | 0 | — |
case-01 | fail→pass | 36,272 | 29,597 | -18% | 1 | 1 | 0% | 6,250 | 7,215 | +15% | 0 | 0 | — |
case-02 | fail→pass | 53,458 | 39,543 | -26% | 1 | 1 | 0% | 6,220 | 8,343 | +34% | 0 | 0 | — |
case-03 | fail→fail | 35,500 | 40,014 | +13% | 1 | 1 | 0% | 5,619 | 8,357 | +49% | 0 | 0 | — |
case-04 | fail→pass | 14,524 | 19,639 | +35% | 1 | 1 | 0% | 2,248 | 5,207 | +132% | 0 | 0 | — |
case-05 | fail→fail | 15,528 | 15,737 | +1% | 1 | 1 | 0% | 2,177 | 4,287 | +97% | 0 | 0 | — |
case-06 | pass→pass | 17,536 | 20,293 | +16% | 1 | 1 | 0% | 2,738 | 5,327 | +95% | 0 | 0 | — |
case-07 | pass→pass | 14,683 | 16,019 | +9% | 1 | 1 | 0% | 2,262 | 4,383 | +94% | 0 | 0 | — |
case-08 | pass→pass | 13,829 | 15,988 | +16% | 1 | 1 | 0% | 1,988 | 4,584 | +131% | 0 | 0 | — |
case-09 | pass→pass | 12,268 | 10,532 | -14% | 1 | 1 | 0% | 1,971 | 4,020 | +104% | 0 | 0 | — |
case-10 | pass→pass | 16,398 | 16,679 | +2% | 1 | 1 | 0% | 2,611 | 4,916 | +88% | 0 | 0 | — |
case-11 | pass→pass | 15,603 | 17,988 | +15% | 1 | 1 | 0% | 2,481 | 4,970 | +100% | 0 | 0 | — |
case-12 | pass→pass | 12,037 | 17,276 | +44% | 1 | 1 | 0% | 1,966 | 4,874 | +148% | 0 | 0 | — |
case-13 | pass→pass | 15,376 | 17,928 | +17% | 1 | 1 | 0% | 2,344 | 4,874 | +108% | 0 | 0 | — |
case-14 | pass→pass | 12,361 | 13,706 | +11% | 1 | 1 | 0% | 1,970 | 4,254 | +116% | 0 | 0 | — |
case-15 | pass→pass | 8,520 | 9,039 | +6% | 1 | 1 | 0% | 1,512 | 3,587 | +137% | 0 | 0 | — |
case-16 | pass→pass | 17,786 | 22,201 | +25% | 1 | 1 | 0% | 2,907 | 5,422 | +87% | 0 | 0 | — |
case-18 | fail→pass | 23,155 | 17,704 | -24% | 1 | 1 | 0% | 3,688 | 5,269 | +43% | 0 | 0 | — |
case-19 | pass→pass | 14,627 | 13,681 | -6% | 1 | 1 | 0% | 2,011 | 3,999 | +99% | 0 | 0 | — |
case-20 | pass→pass | 10,018 | 19,641 | +96% | 1 | 1 | 0% | 1,912 | 5,551 | +190% | 0 | 0 | — |
case-21 | pass→pass | 15,581 | 19,070 | +22% | 1 | 1 | 0% | 2,516 | 5,016 | +99% | 0 | 0 | — |
case-22 | pass→pass | 20,468 | 16,248 | -21% | 1 | 1 | 0% | 2,922 | 4,596 | +57% | 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 +18 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.