0 Fork 0

Files

T

Evan Hu ed595230cf feat(hands): improve 6 lower-scoring hands — system prompts and SKILL.md depth

- browser: 5→7 phases, SPA detection, error recovery decision tree, 3 new settings
- strategist: framework integration methodology, 7 anti-patterns, uncertainty quantification
- lead: remove clip language, add BANT/MEDDIC qualification, 3 new settings + CRM export
- researcher: CRAAP→CRAAP+, 7-step conflict resolution, 6-item cognitive bias audit
- collector: concrete change classification (structural/content/metadata), 5-factor scoring, 2 new settings
- apitester: OWASP Top 10 checklist, 4 load test profiles, contract testing phase, GraphQL/Webhook patterns

2026-03-23 00:31:13 +09:00

HAND.toml

feat(hands): improve 6 lower-scoring hands — system prompts and SKILL.md depth

2026-03-23 00:31:13 +09:00

README.md

docs: add README for every subdirectory

2026-03-21 02:27:34 +09:00

SKILL.md

feat(hands): improve 6 lower-scoring hands — system prompts and SKILL.md depth

2026-03-23 00:31:13 +09:00

README.md

Researcher Hand

Autonomous deep researcher -- exhaustive investigation, cross-referencing, fact-checking, and structured reports.

Configuration

Field	Value
Category	`productivity`
Agent	`researcher-hand`
Routing	`deep research`, `systematic review`, `landscape analysis`, `exhaustive investigation`

Integrations

None required.

Settings

Research Depth -- quick (5-10 sources), thorough (20-30), exhaustive (50+)
Output Style -- brief, detailed, academic, executive
Source Verification -- Cross-check claims across multiple sources (default: on)
Max Sources -- 10, 30, 50, unlimited
Auto Follow-Up -- Research follow-up questions discovered during investigation (default: on)
Save Research Log -- Save detailed search queries and source notes (default: off)
Citation Style -- inline_url, footnotes, academic_apa, numbered
Language -- english, spanish, french, german, chinese, japanese, auto

Usage

librefang hand run researcher