Tool tree

secureflow-review-room.daniel-ca-pe207.chatgpt.site

Developer ToolsTools verifiedOct 26, 2026

An agent-native security review workspace where AI investigates structured evidence and a human owns every final decision.

list_candidatesAnswer

List the structured candidates in this authorized review case. Candidates are not confirmed vulnerabilities.

View tool JSON

{
  "name": "list_candidates",
  "kind": "answer",
  "impl": "imperative",
  "description": "List the structured candidates in this authorized review case. Candidates are not confirmed vulnerabilities.",
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false
  }
}

inspect_evidenceAction

Select a candidate in the visible review room and return its evidence boundary, source, visible guard, and sink.

View tool JSON

{
  "name": "inspect_evidence",
  "kind": "act",
  "impl": "imperative",
  "description": "Select a candidate in the visible review room and return its evidence boundary, source, visible guard, and sink.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "candidateId": {
        "type": "string",
        "enum": [
          "AUTHZ-014",
          "WEBHOOK-009",
          "CACHE-003"
        ]
      }
    },
    "required": [
      "candidateId"
    ],
    "additionalProperties": false
  }
}

compare_revisionAction

Select a candidate and return the supplied revision note without claiming exploitability.

View tool JSON

{
  "name": "compare_revision",
  "kind": "act",
  "impl": "imperative",
  "description": "Select a candidate and return the supplied revision note without claiming exploitability.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "candidateId": {
        "type": "string",
        "enum": [
          "AUTHZ-014",
          "WEBHOOK-009",
          "CACHE-003"
        ]
      }
    },
    "required": [
      "candidateId"
    ],
    "additionalProperties": false
  }
}

draft_hardeningAction

Return an evidence-bound hardening direction for a candidate. This does not confirm a vulnerability or modify code.

View tool JSON

{
  "name": "draft_hardening",
  "kind": "act",
  "impl": "imperative",
  "description": "Return an evidence-bound hardening direction for a candidate. This does not confirm a vulnerability or modify code.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "candidateId": {
        "type": "string",
        "enum": [
          "AUTHZ-014",
          "WEBHOOK-009",
          "CACHE-003"
        ]
      }
    },
    "required": [
      "candidateId"
    ],
    "additionalProperties": false
  }
}

stage_agent_recommendationAction

Stage a provisional recommendation and rationale in the visible human decision form. This cannot record or finalize a security decision.

View tool JSON

{
  "name": "stage_agent_recommendation",
  "kind": "act",
  "impl": "imperative",
  "description": "Stage a provisional recommendation and rationale in the visible human decision form. This cannot record or finalize a security decision.",
  "inputSchema": {
    "type": "object",
    "properties": {
      "candidateId": {
        "type": "string",
        "enum": [
          "AUTHZ-014",
          "WEBHOOK-009",
          "CACHE-003"
        ]
      }
    },
    "required": [
      "candidateId"
    ],
    "additionalProperties": false
  }
}