@arbol-ai/cli 2.8.0 → 2.9.0

package/dist/index.js

@@ -16859,6 +16859,121 @@ var commands = {
     confirmation: null,
     hidden: false
   },
+  "raiz eval-cases create": {
+    operation: "raiz.evalCases.create",
+    sdkPath: [
+      "raiz",
+      "evalCases",
+      "create"
+    ],
+    pathField: null,
+    fields: [
+      {
+        name: "conversationId",
+        flag: "conversation-id",
+        kind: "string",
+        required: true,
+        description: "Conversation id."
+      },
+      {
+        name: "replyMessageId",
+        flag: "reply-message-id",
+        kind: "string",
+        required: true,
+        description: "Message id."
+      },
+      {
+        name: "expectation",
+        flag: "expectation",
+        kind: "string",
+        required: true,
+        description: "What the Agent should have done instead, for whoever replays the case: a short rubric, never the contact's personal data."
+      },
+      {
+        name: "preferredResponse",
+        flag: "preferred-response",
+        kind: "string",
+        required: true,
+        description: "The team's preferred reply or procedure. Private evidence, never automatically instructions."
+      },
+      {
+        name: "feedbackKind",
+        flag: "feedback-kind",
+        kind: "string",
+        required: true,
+        description: "",
+        choices: [
+          "incorrect",
+          "simpler",
+          "incomplete"
+        ]
+      }
+    ],
+    inputSchema: {
+      additionalProperties: false,
+      type: "object",
+      required: [
+        "conversationId",
+        "replyMessageId",
+        "expectation",
+        "preferredResponse",
+        "feedbackKind"
+      ],
+      properties: {
+        conversationId: {
+          description: "Conversation id.",
+          examples: [
+            "cnv_01m45h8q2rf3t9w6y7z0b1c2d3"
+          ],
+          pattern: "^cnv_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+          type: "string"
+        },
+        replyMessageId: {
+          description: "Message id.",
+          examples: [
+            "msg_01m45h8q2rf3t9w6y7z0b1c2d3"
+          ],
+          pattern: "^msg_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+          type: "string"
+        },
+        expectation: {
+          description: "What the Agent should have done instead, for whoever replays the case: a short rubric, never the contact's personal data.",
+          maxLength: 4000,
+          minLength: 1,
+          pattern: "\\S",
+          type: "string"
+        },
+        preferredResponse: {
+          minLength: 1,
+          maxLength: 4000,
+          pattern: "\\S",
+          description: "The team's preferred reply or procedure. Private evidence, never automatically instructions.",
+          type: "string"
+        },
+        feedbackKind: {
+          anyOf: [
+            {
+              const: "incorrect",
+              type: "string"
+            },
+            {
+              const: "simpler",
+              type: "string"
+            },
+            {
+              const: "incomplete",
+              type: "string"
+            }
+          ]
+        }
+      }
+    },
+    title: "Record Agent feedback",
+    description: "Record internal feedback on one completed WhatsApp Agent reply: what should be preserved or changed, a preferred response, and whether it was incorrect, incomplete or could be simpler. Requires conversation write and Agent read access. Does not contact anyone or change live instructions. Repeating creation returns the existing authored case; edit it with updateEvalCase and expectedVersion.",
+    effect: "write",
+    confirmation: null,
+    hidden: false
+  },
   "raiz eval-cases list": {
     operation: "raiz.evalCases.list",
     sdkPath: [
@@ -16913,6 +17028,25 @@ var commands = {
           "kept",
           "dismissed"
         ]
+      },
+      {
+        name: "replyMessageId",
+        flag: "reply-message-id",
+        kind: "string",
+        required: false,
+        description: "Message id."
+      },
+      {
+        name: "source",
+        flag: "source",
+        kind: "string",
+        required: false,
+        description: "",
+        choices: [
+          "correction",
+          "sample",
+          "authored"
+        ]
       }
     ],
     inputSchema: {
@@ -16967,12 +17101,36 @@ var commands = {
               type: "string"
             }
           ]
+        },
+        replyMessageId: {
+          description: "Message id.",
+          examples: [
+            "msg_01m45h8q2rf3t9w6y7z0b1c2d3"
+          ],
+          pattern: "^msg_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+          type: "string"
+        },
+        source: {
+          anyOf: [
+            {
+              const: "correction",
+              type: "string"
+            },
+            {
+              const: "sample",
+              type: "string"
+            },
+            {
+              const: "authored",
+              type: "string"
+            }
+          ]
         }
       },
       additionalProperties: false
     },
     title: "List evaluation cases",
-    description: "The Agents' evaluation cases, newest first: each is a WhatsApp reply the model wrote after which a person from the business wrote in the conversation (`correction`), with the Agent version that wrote it. Filter by `agentId` or `status` (`open` not reviewed yet, `kept`, `dismissed`). Read the conversation with its own operations to judge a case. Not an AI tool.",
+    description: "The Agents' evaluation cases, newest first: completed WhatsApp replies with explicit internal team feedback (`authored`) or a following team message (`correction`), and the Agent version that wrote each reply. Filter by agentId, replyMessageId, source or status (`open`, `kept`, `dismissed`). Requires conversation read authority as well as Agent read authority. Read the conversation with its own operations to judge a case. Not an AI tool.",
     effect: "read",
     confirmation: null,
     hidden: false
@@ -16993,6 +17151,32 @@ var commands = {
         required: true,
         description: "Evaluation case id."
       },
+      {
+        name: "expectedVersion",
+        flag: "expected-version",
+        kind: "integer",
+        required: false,
+        description: "The `version` you read. When present, the change applies only if nobody changed the resource since; otherwise it fails with VERSION_CONFLICT."
+      },
+      {
+        name: "preferredResponse",
+        flag: "preferred-response",
+        kind: "string",
+        required: false,
+        description: ""
+      },
+      {
+        name: "feedbackKind",
+        flag: "feedback-kind",
+        kind: "string",
+        required: false,
+        description: "",
+        choices: [
+          "incorrect",
+          "simpler",
+          "incomplete"
+        ]
+      },
       {
         name: "expectation",
         flag: "expectation",
@@ -17028,12 +17212,48 @@ var commands = {
           pattern: "^evc_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
           type: "string"
         },
+        expectedVersion: {
+          description: "The `version` you read. When present, the change applies only if nobody changed the resource since; otherwise it fails with VERSION_CONFLICT.",
+          minimum: 1,
+          type: "integer"
+        },
+        preferredResponse: {
+          anyOf: [
+            {
+              minLength: 1,
+              maxLength: 4000,
+              pattern: "\\S",
+              description: "The team's preferred reply or procedure. Private evidence, never automatically instructions.",
+              type: "string"
+            },
+            {
+              type: "null"
+            }
+          ]
+        },
+        feedbackKind: {
+          anyOf: [
+            {
+              const: "incorrect",
+              type: "string"
+            },
+            {
+              const: "simpler",
+              type: "string"
+            },
+            {
+              const: "incomplete",
+              type: "string"
+            }
+          ]
+        },
         expectation: {
           anyOf: [
             {
               description: "What the Agent should have done instead, for whoever replays the case: a short rubric, never the contact's personal data.",
               maxLength: 4000,
               minLength: 1,
+              pattern: "\\S",
               type: "string"
             },
             {
@@ -17082,6 +17302,13 @@ var commands = {
         required: true,
         description: "Agent id."
       },
+      {
+        name: "feedbackCaseId",
+        flag: "feedback-case-id",
+        kind: "string",
+        required: false,
+        description: "Evaluation case id."
+      },
       {
         name: "caseIds",
         flag: "case-ids",
@@ -17212,6 +17439,14 @@ var commands = {
             }
           }
         },
+        feedbackCaseId: {
+          description: "Evaluation case id.",
+          examples: [
+            "evc_01m45h8q2rf3t9w6y7z0b1c2d3"
+          ],
+          pattern: "^evc_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+          type: "string"
+        },
         caseIds: {
           description: "The Agent's kept cases to replay. Omitted: its 50 newest kept cases.",
           maxItems: 50,
@@ -17235,7 +17470,7 @@ var commands = {
       }
     },
     title: "Replay evaluation cases on a candidate",
-    description: "Measure a candidate version of a text Agent on its kept evaluation cases before publishing it: the WhatsApp Agent rewrites each case's reply with the candidate's `instructions`, `skills`, `integrationSkills`, `model` and `reasoningEffort` (omitted ones stay as the live Agent has them), without running any integration tool (each answers with what it answered in the original conversation) and without sending anything; then a judge model compares each new reply with what the team expects. Replays every kept case, or `caseIds`, at most 50. It calls paid models, so it needs `confirmCost: true`. Follow it with `getEvalRun`; with the Agent's `settings.evalGate`, `updateAgent` publishes exactly this content with `evalRunId`. EVAL_CASES_UNAVAILABLE: no kept case to replay. Not an AI tool.",
+    description: "With feedbackCaseId, prepare conditional guidance from that kept correction and test it with the Agent’s newest kept cases, always including the source. Do not combine feedbackCaseId with candidate or caseIds. This does not publish. Requires conversation read access. Otherwise measure a candidate version of a text Agent on its kept evaluation cases before publishing it: the WhatsApp Agent rewrites each case's reply with the candidate's `instructions`, `skills`, `integrationSkills`, `model` and `reasoningEffort` (omitted ones stay as the live Agent has them), without running any integration tool (each answers with what it answered in the original conversation) and without sending anything; then a judge model compares each new reply with what the team expects. Replays every kept case, or `caseIds`, at most 50. It calls paid models, so it needs `confirmCost: true`. Follow it with `getEvalRun`; with the Agent's `settings.evalGate`, `updateAgent` publishes exactly this content with `evalRunId`. EVAL_CASES_UNAVAILABLE: no kept case to replay. Not an AI tool.",
     effect: "write",
     confirmation: null,
     hidden: false
@@ -45901,6 +46136,10 @@ var schemaComponents = {
       "correctionMessageId",
       "createdAt",
       "expectation",
+      "preferredResponse",
+      "feedbackKind",
+      "version",
+      "latestRunId",
       "id",
       "replyMessageId",
       "source",
@@ -45965,6 +46204,58 @@ var schemaComponents = {
           }
         ]
       },
+      preferredResponse: {
+        anyOf: [
+          {
+            type: "string"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      feedbackKind: {
+        anyOf: [
+          {
+            anyOf: [
+              {
+                const: "incorrect",
+                type: "string"
+              },
+              {
+                const: "simpler",
+                type: "string"
+              },
+              {
+                const: "incomplete",
+                type: "string"
+              }
+            ]
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      version: {
+        minimum: 1,
+        type: "integer"
+      },
+      latestRunId: {
+        anyOf: [
+          {
+            description: "The latest candidate prepared from this correction.",
+            examples: [
+              "evr_01m45h8q2rf3t9w6y7z0b1c2d3"
+            ],
+            pattern: "^evr_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+            type: "string"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
       id: {
         description: "Evaluation case id.",
         examples: [
@@ -45982,7 +46273,7 @@ var schemaComponents = {
         type: "string"
       },
       source: {
-        description: "`correction`: a person wrote right after the reply. `sample` and `authored` are reserved for sampled and hand-written cases.",
+        description: "`correction`: a person wrote right after the reply. `authored`: explicit internal team feedback. `sample`: a sampled reply.",
         anyOf: [
           {
             const: "correction",
@@ -46054,8 +46345,13 @@ var schemaComponents = {
     type: "object",
     required: [
       "agentId",
+      "baseVersion",
+      "feedbackCaseVersion",
       "baseVersionId",
       "caseIds",
+      "feedbackCaseId",
+      "proposal",
+      "proposedInstructions",
       "contentSha256",
       "createdAt",
       "endedAt",
@@ -46077,6 +46373,28 @@ var schemaComponents = {
         pattern: "^agt_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
         type: "string"
       },
+      baseVersion: {
+        anyOf: [
+          {
+            minimum: 1,
+            type: "integer"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      feedbackCaseVersion: {
+        anyOf: [
+          {
+            minimum: 1,
+            type: "integer"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
       baseVersionId: {
         anyOf: [
           {
@@ -46100,6 +46418,77 @@ var schemaComponents = {
           type: "string"
         }
       },
+      feedbackCaseId: {
+        anyOf: [
+          {
+            description: "Evaluation case id.",
+            examples: [
+              "evc_01m45h8q2rf3t9w6y7z0b1c2d3"
+            ],
+            pattern: "^evc_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+            type: "string"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      proposal: {
+        anyOf: [
+          {
+            additionalProperties: false,
+            type: "object",
+            required: [
+              "appliesWhen",
+              "guidance",
+              "exceptions"
+            ],
+            properties: {
+              replaceText: {
+                anyOf: [
+                  {
+                    minLength: 1,
+                    maxLength: 2000,
+                    type: "string"
+                  },
+                  {
+                    type: "null"
+                  }
+                ]
+              },
+              appliesWhen: {
+                minLength: 1,
+                maxLength: 600,
+                type: "string"
+              },
+              guidance: {
+                minLength: 1,
+                maxLength: 1200,
+                type: "string"
+              },
+              exceptions: {
+                minLength: 1,
+                maxLength: 600,
+                type: "string"
+              }
+            }
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      proposedInstructions: {
+        anyOf: [
+          {
+            description: "For a feedback proposal, the exact candidate instructions to review and send to updateAgent with evalRunId and expectedVersion. No live Agent has changed.",
+            type: "string"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
       contentSha256: {
         description: "SHA-256 of the candidate's content (instructions, skills, connection skills, model, reasoning effort): what `settings.evalGate` compares with an update's.",
         pattern: "^[a-f0-9]{64}$",
@@ -49070,6 +49459,35 @@ var schemaComponents = {
               }
             }
           },
+          {
+            additionalProperties: false,
+            type: "object",
+            required: [
+              "code",
+              "message",
+              "retryable",
+              "data"
+            ],
+            properties: {
+              code: {
+                const: "FEEDBACK_PROPOSAL_UNAVAILABLE",
+                type: "string"
+              },
+              message: {
+                const: "A usable improvement could not be prepared. The Agent has not changed. Try again later.",
+                type: "string"
+              },
+              retryable: {
+                const: true,
+                type: "boolean"
+              },
+              data: {
+                additionalProperties: false,
+                type: "object",
+                properties: {}
+              }
+            }
+          },
           {
             additionalProperties: false,
             type: "object",
@@ -68331,10 +68749,93 @@ var schemaComponents = {
             type: "string"
           }
         ]
+      },
+      replyMessageId: {
+        description: "Message id.",
+        examples: [
+          "msg_01m45h8q2rf3t9w6y7z0b1c2d3"
+        ],
+        pattern: "^msg_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+        type: "string"
+      },
+      source: {
+        anyOf: [
+          {
+            const: "correction",
+            type: "string"
+          },
+          {
+            const: "sample",
+            type: "string"
+          },
+          {
+            const: "authored",
+            type: "string"
+          }
+        ]
       }
     },
     additionalProperties: false
   },
+  RaizEvalCasesCreateInput: {
+    additionalProperties: false,
+    type: "object",
+    required: [
+      "conversationId",
+      "replyMessageId",
+      "expectation",
+      "preferredResponse",
+      "feedbackKind"
+    ],
+    properties: {
+      conversationId: {
+        description: "Conversation id.",
+        examples: [
+          "cnv_01m45h8q2rf3t9w6y7z0b1c2d3"
+        ],
+        pattern: "^cnv_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+        type: "string"
+      },
+      replyMessageId: {
+        description: "Message id.",
+        examples: [
+          "msg_01m45h8q2rf3t9w6y7z0b1c2d3"
+        ],
+        pattern: "^msg_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+        type: "string"
+      },
+      expectation: {
+        description: "What the Agent should have done instead, for whoever replays the case: a short rubric, never the contact's personal data.",
+        maxLength: 4000,
+        minLength: 1,
+        pattern: "\\S",
+        type: "string"
+      },
+      preferredResponse: {
+        minLength: 1,
+        maxLength: 4000,
+        pattern: "\\S",
+        description: "The team's preferred reply or procedure. Private evidence, never automatically instructions.",
+        type: "string"
+      },
+      feedbackKind: {
+        anyOf: [
+          {
+            const: "incorrect",
+            type: "string"
+          },
+          {
+            const: "simpler",
+            type: "string"
+          },
+          {
+            const: "incomplete",
+            type: "string"
+          }
+        ]
+      }
+    }
+  },
   RaizEvalCasesUpdateInput: {
     additionalProperties: false,
     type: "object",
@@ -68350,12 +68851,48 @@ var schemaComponents = {
         pattern: "^evc_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
         type: "string"
       },
+      expectedVersion: {
+        description: "The `version` you read. When present, the change applies only if nobody changed the resource since; otherwise it fails with VERSION_CONFLICT.",
+        minimum: 1,
+        type: "integer"
+      },
+      preferredResponse: {
+        anyOf: [
+          {
+            minLength: 1,
+            maxLength: 4000,
+            pattern: "\\S",
+            description: "The team's preferred reply or procedure. Private evidence, never automatically instructions.",
+            type: "string"
+          },
+          {
+            type: "null"
+          }
+        ]
+      },
+      feedbackKind: {
+        anyOf: [
+          {
+            const: "incorrect",
+            type: "string"
+          },
+          {
+            const: "simpler",
+            type: "string"
+          },
+          {
+            const: "incomplete",
+            type: "string"
+          }
+        ]
+      },
       expectation: {
         anyOf: [
           {
             description: "What the Agent should have done instead, for whoever replays the case: a short rubric, never the contact's personal data.",
             maxLength: 4000,
             minLength: 1,
+            pattern: "\\S",
             type: "string"
           },
           {
@@ -68494,6 +69031,14 @@ var schemaComponents = {
           }
         }
       },
+      feedbackCaseId: {
+        description: "Evaluation case id.",
+        examples: [
+          "evc_01m45h8q2rf3t9w6y7z0b1c2d3"
+        ],
+        pattern: "^evc_(?:[0-7][0-9a-hjkmnp-tv-z]{25}|(?:[a-f0-9]{8}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{4}-[a-f0-9]{12}|[a-f0-9]{32}))$",
+        type: "string"
+      },
       caseIds: {
         description: "The Agent's kept cases to replay. Omitted: its 50 newest kept cases.",
         maxItems: 50,
@@ -73249,9 +73794,9 @@ var schemaComponents = {
 // src/generated/generation.json
 var generation_default = {
   formatVersion: 6,
-  release: "1.20261023.34948",
-  openapiSha256: "536c2fe460227332df9c3188c88b557917ec321ccc30839c4bbd4155eef67c2d",
-  openapiContentSha256: "5e0595dc6bae6f9f6ac64d2a632d1b1d6bf0bfbdbd2bb939a80407b38f69f96a",
+  release: "1.20261023.60554",
+  openapiSha256: "86323054a3e408946c70daf18480c82b7084c3e38c73e469d0d34227bf10a117",
+  openapiContentSha256: "c05fa12d8b1bed4efd2d1e6a0e29a7dae23ee0e665b8117d08c7dc1d458876d4",
   forge: {
     repository: "https://github.com/cloudflare/forge",
     commit: "7f86447210b7bea3779bfdab3cd5655eade0eadb",
@@ -73264,7 +73809,7 @@ var generation_default = {
   outputs: [
     {
       path: "src/generated/commands.ts",
-      sha256: "6584cbdbc4dc8700562d4b5e3a895120af47200163f43a75385a70d319241122"
+      sha256: "a06f367dba10667942695d2565afd2bb43fa1ac72279cec15effe052d47e46b0"
     }
   ]
 };
@@ -73619,7 +74164,7 @@ ${name}`);
 }
 
 // src/index.ts
-var VERSION = "2.8.0";
+var VERSION = "2.9.0";
 var RELEASE = generation_default.release;
 var SDK_PATHS = new Map(Object.values(commands).map((command) => [command.operation, command.sdkPath]));
 var SERVICES = [...new Set(Object.keys(commands).map((name) => name.split(" ")[0]))].toSorted();

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@arbol-ai/cli",
-  "version": "2.8.0",
+  "version": "2.9.0",
   "description": "The Arbol CLI: the Arbol API from your terminal, with commands generated by Cloudflare Forge from the public OpenAPI.",
   "keywords": [
     "arbol",
@@ -44,7 +44,7 @@
     "verify": "bun run typecheck && bun run test && bun run build"
   },
   "dependencies": {
-    "@arbol-ai/sdk": "2.9.0",
+    "@arbol-ai/sdk": "2.10.0",
     "ajv": "8.20.0",
     "ajv-formats": "3.0.1",
     "oauth4webapi": "3.8.8"
@@ -59,8 +59,8 @@
   "packageManager": "bun@1.4.3",
   "arbol": {
     "openapi": {
-      "release": "1.20261023.34948",
-      "sha256": "5e0595dc6bae6f9f6ac64d2a632d1b1d6bf0bfbdbd2bb939a80407b38f69f96a"
+      "release": "1.20261023.60554",
+      "sha256": "c05fa12d8b1bed4efd2d1e6a0e29a7dae23ee0e665b8117d08c7dc1d458876d4"
     }
   }
 }