Evidence Research OS Search Methodologies

New Methodology

Starts from a minimal valid research pipeline — edit, save, validation runs server-side.

Methodology YAML

id: my-methodology-v1
name: "My Methodology"
description: "Describe what this pipeline does"
is_default: false
compatible_modes: [research]
workflow:
  stages:
    - {id: trigger_classifier, node: trigger_classifier, route: route_classifier}
    - {id: plan, node: plan}
    - {id: independent_first_pass, node: independent_first_pass, roles: [scientist, investigator, skeptic]}
    - {id: evidence_extraction, node: evidence_extraction}
    - {id: conflict_detection, node: conflict_detection, route: route_conflict}
    - {id: targeted_research, node: targeted_research}
    - {id: adversarial_review, node: adversarial_review}
    - {id: evidence_adjudication, node: evidence_adjudication}
    - {id: synthesis, node: synthesis}
    - {id: citation_audit, node: citation_audit, route: route_audit}
    - {id: targeted_repair, node: targeted_repair}
    - {id: human_checkpoint, node: human_checkpoint, interrupt: true}
    - {id: final_output, node: final_output}
tools:
  enabled: [search_web, fetch_url, fetch_pdf, grep_project, keyword_search, semantic_search, citation_verify, store_source, retrieve_evidence]
prompts:
  set: role-prompts/v1
  overrides: {}
skills:
  scientist: [hypothesis-decomposition]
  investigator: [source-retrieval, contradiction-search]
  skeptic: [adversarial-review-research]
  judge: [evidence-adjudication]
models:
  scientist: ""
  investigator: ""
  skeptic: ""
  judge: ""
  ideator: ""
  auditor: ""
budget_defaults:
  max_model_calls: 50
  max_research_rounds: 5

Stage preview (approximate)

No stages detected.