verdiqt-web.onrender.com

{
  "name": "start_validation",
  "kind": "act",
  "impl": "imperative",
  "description": "START HERE to validate a SaaS idea. Provide either idea_text describing the idea, or repo_url pointing to a public GitHub repository. Returns a run_id. WORKFLOW: 1) call this, 2) poll get_validation_status every few seconds until status is COMPLETE (about 30s), 3) call get_verdict for the final BUILD/PIVOT/KILL ruling, 4) optionally get_evidence and get_next_step. Report the verdict, score, and next step to the user.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "idea_text": {
        "type": "string",
        "minLength": 1,
        "maxLength": 2000
      },
      "repo_url": {
        "type": "string",
        "format": "uri"
      }
    },
    "oneOf": [
      {
        "properties": {
          "idea_text": {},
          "repo_url": false
        },
        "required": [
          "idea_text"
        ]
      },
      {
        "properties": {
          "idea_text": false,
          "repo_url": {}
        },
        "required": [
          "repo_url"
        ]
      }
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "get_validation_status",
  "kind": "answer",
  "impl": "imperative",
  "description": "Poll a running trial by run_id. Statuses: QUEUED, NORMALIZING, GATHERING, CLASSIFYING, SCORING, COMPLETE, FAILED. Keep polling every few seconds while not terminal; when COMPLETE call get_verdict; when FAILED read error_code and tell the user why. Also returns evidence_count, per-source research states, the normalized case_file, pending approvals, and recent human actions — re-read after the human changes anything.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      }
    },
    "required": [
      "run_id"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "get_evidence",
  "kind": "answer",
  "impl": "imperative",
  "description": "List evidence gathered for a trial. Optionally filter by dimension or source. Includes the human's pinned or rejected state per item.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      },
      "dimension": {
        "type": "string",
        "enum": [
          "PROBLEM_SEVERITY",
          "DEMAND_SIGNALS",
          "COMPETITION",
          "MONETIZATION",
          "DISTRIBUTION",
          "BUILD_COST"
        ]
      },
      "source": {
        "type": "string",
        "enum": [
          "WEB_SEARCH",
          "REDDIT",
          "HACKERNEWS",
          "PRODUCT_HUNT",
          "GITHUB"
        ]
      }
    },
    "required": [
      "run_id"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "request_deep_scan",
  "kind": "act",
  "impl": "imperative",
  "description": "Request a deeper evidence scan for one dimension of a trial. This queues an approval card in the page UI; the human must click Approve before the scan runs. Poll get_validation_status to see when it completes.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      },
      "dimension": {
        "type": "string",
        "enum": [
          "PROBLEM_SEVERITY",
          "DEMAND_SIGNALS",
          "COMPETITION",
          "MONETIZATION",
          "DISTRIBUTION",
          "BUILD_COST"
        ]
      },
      "reason": {
        "type": "string",
        "maxLength": 300
      }
    },
    "required": [
      "run_id",
      "dimension"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "get_verdict",
  "kind": "answer",
  "impl": "imperative",
  "description": "Get the final ruling for a COMPLETE trial: composite score 0-100, BUILD (70+) / PIVOT (40-69) / KILL (<40), six dimension scores with rationales and evidence citations, the bench's written opinion (next_step.bench_opinion), and the single recommended next step. Summarize the verdict, the bench opinion, and the next step for the user; cite evidence via get_evidence when asked for proof.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      }
    },
    "required": [
      "run_id"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "refine_idea",
  "kind": "act",
  "impl": "imperative",
  "description": "Re-run validation on a pivoted version of an existing idea. Creates a new linked trial so the two can be compared with compare_ideas.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      },
      "pivot_text": {
        "type": "string",
        "minLength": 1,
        "maxLength": 2000
      }
    },
    "required": [
      "run_id",
      "pivot_text"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "compare_ideas",
  "kind": "act",
  "impl": "imperative",
  "description": "Compare two or more completed trials side by side across all six dimensions.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_ids": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "minItems": 2,
        "maxItems": 5,
        "uniqueItems": true
      }
    },
    "required": [
      "run_ids"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "get_next_step",
  "kind": "answer",
  "impl": "imperative",
  "description": "Get the single recommended next validation action for a completed trial, chosen from Verdiqt's validation playbook.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "run_id": {
        "type": "string"
      }
    },
    "required": [
      "run_id"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "analyze_repo",
  "kind": "act",
  "impl": "imperative",
  "description": "Start a validation trial from any public GitHub repository (owner/repo). The court reads the repo's README and metadata to infer the idea, then runs the full evidence trial. Poll get_validation_status until COMPLETE, then call get_verdict.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "repo_full_name": {
        "type": "string",
        "minLength": 3,
        "maxLength": 200,
        "pattern": "^[^/\\s]+/[^/\\s]+$"
      }
    },
    "required": [
      "repo_full_name"
    ],
    "additionalProperties": false
  },
  "page": "/"
}

{
  "name": "search_knowledge",
  "kind": "answer",
  "impl": "imperative",
  "description": "Search Verdiqt's marketing and validation knowledge base for frameworks and heuristics, for example offer design, demand signals, pricing, distribution.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "minLength": 1,
        "maxLength": 200
      },
      "tags": {
        "type": "array",
        "items": {
          "type": "string",
          "enum": [
            "PROBLEM_SEVERITY",
            "DEMAND_SIGNALS",
            "COMPETITION",
            "MONETIZATION",
            "DISTRIBUTION",
            "BUILD_COST"
          ]
        },
        "maxItems": 6,
        "uniqueItems": true
      },
      "limit": {
        "type": "integer",
        "minimum": 1,
        "maximum": 10,
        "default": 6
      }
    },
    "required": [
      "query"
    ],
    "additionalProperties": false
  },
  "page": "/"
}