From searchnet
Claim verification with dual-track evidence collection and credibility scoring. Best for "is it true that X?" and fact-checking questions.
How this skill is triggered — by the user, by Claude, or both
Slash command
/searchnet:verificationThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
You are the verification search orchestrator. This pipeline checks claims:
You are the verification search orchestrator. This pipeline checks claims: "is it true that X?", "does Y actually do Z?", "verify this statement."
Before starting the search, present the user with:
Ask: "This will launch ~N agents (depends on claim count). Proceed?"
Only continue to Step 1 after user confirms.
workers/ and leaders/ subdirectoriessearch-plan.mdFor each claim, launch 2 worker groups in parallel (background). Use
searchnet:haiku-searcher.
Each worker writes to workers/{claim}-{support|contradict}-{n}.md.
After all workers complete, launch one searchnet:opus-searcher leader per
claim in parallel (background).
Each leader:
leaders/leader-{claim}.mdRead all leader assessments. Produce the final verification report to
final-report.md.
Launch the searchnet:verifier agent on final-report.md.
npx claudepluginhub pablolion/searchnet --plugin searchnetVerifies claims using adversarial search: finds supporting and contradicting evidence, assesses source quality by tiers, and issues structured verdict with confidence. Counters LLM confirmation bias.
Verifies factual claims by isolating assertions, gathering primary-source evidence, and rating accuracy with confidence levels. Useful for checking stats, quotes, and viral claims.
Triangulates truth from diverse sources by cross-validating claims, auditing source credibility, and producing a verification matrix. Useful for fact-checking and investigating contradictory information.