Agent safely updates a task document

update-task-document

The agent reads the current document revision before making an optimistic update.

Prompt

Read the current eval-notes document, then update it by appending a second line: Next assertion: context before action. Preserve the existing text and use the current revision id as baseRevisionId.

Fixture

../fixtures/companies/orientation.json

Authority

{
  "actorId": "actor-engineer",
  "capabilities": [],
  "explicitClaims": [],
  "taskId": "task-orientation"
}

Current results

Model	Configuration	Result	Estimated cost	Run time
claude-sonnet-5	live-acpx-0.13.1-claude-acp-0.70.0-sonnet-5	INFRA	unknown	2026-09-06T20:52:24.867Z
claude-sonnet-5	live-claude-managed-agents-2026-04-01-sonnet-5	PASS	$0.025764 estimated	2026-09-06T20:52:39.849Z
global.anthropic.claude-sonnet-4-6	live-aws-agentcore-global-anthropic-claude-sonnet-4-6	INFRA	unknown	2026-09-06T20:53:19.227Z
gpt-5.4-mini	live-codex-pinned	INFRA	unknown	unknown
gpt-5.6-sol	live-acpx-0.13.1-codex-acp-1.6.2-control	Not run	—	—
openrouter/anthropic/claude-sonnet-5	live-opencode-1.18.17-openrouter-claude-sonnet-5	PASS	$0.020107 estimated	2026-09-06T20:54:30.213Z
openrouter/deepseek/deepseek-v4-flash-0731	live-opencode-1.18.17-openrouter-deepseek-v4-flash-0731	PASS	$0.000585 estimated	2026-09-06T20:57:56.445Z
openrouter/google/gemini-3.8-flash	live-opencode-1.18.17-openrouter-gemini-3.8-flash	PASS	$0.018065 estimated	2026-09-06T20:55:26.542Z
openrouter/openai/gpt-6-astra	live-opencode-1.18.17-openrouter-gpt-6-astra	PASS	$0.049884 estimated	2026-09-06T20:56:55.780Z
openrouter/qwen/qwen3.8-max-0902	live-opencode-1.18.17-openrouter-qwen3.8-max-0902	PASS	$0.016284 estimated	2026-09-06T20:57:08.760Z
openrouter/z-ai/glm-5.3	live-opencode-1.18.17-openrouter-glm-5.3	PASS	$0.011695 estimated	2026-09-06T20:55:37.019Z

Assertions

read-document

semantic_operation

{
  "id": "read-document",
  "kind": "semantic_operation",
  "minCalls": 1,
  "operationId": "read_document",
  "requireSuccess": true
}

write-document

semantic_operation

{
  "id": "write-document",
  "kind": "semantic_operation",
  "maxCalls": 1,
  "minCalls": 1,
  "operationId": "write_document",
  "requireSuccess": true
}

read-before-write

operation_order

{
  "after": "write_document",
  "before": "read_document",
  "id": "read-before-write",
  "kind": "operation_order"
}

no-new-document

state_effect

{
  "collection": "documents",
  "id": "no-new-document",
  "kind": "state_effect",
  "maximumAdded": 0,
  "minimumAdded": 0
}

terminal-completed

terminal

{
  "id": "terminal-completed",
  "kind": "terminal",
  "status": "completed"
}

transcript-complete

transcript_complete

{
  "id": "transcript-complete",
  "kind": "transcript_complete"
}

mock-only

network

{
  "id": "mock-only",
  "kind": "network",
  "realPaperclipRequests": 0
}

Full authored test definition

{
  "authority": {
    "actorId": "actor-engineer",
    "capabilities": [],
    "explicitClaims": [],
    "taskId": "task-orientation"
  },
  "checks": [
    {
      "id": "read-document",
      "kind": "semantic_operation",
      "minCalls": 1,
      "operationId": "read_document",
      "requireSuccess": true
    },
    {
      "id": "write-document",
      "kind": "semantic_operation",
      "maxCalls": 1,
      "minCalls": 1,
      "operationId": "write_document",
      "requireSuccess": true
    },
    {
      "after": "write_document",
      "before": "read_document",
      "id": "read-before-write",
      "kind": "operation_order"
    },
    {
      "collection": "documents",
      "id": "no-new-document",
      "kind": "state_effect",
      "maximumAdded": 0,
      "minimumAdded": 0
    },
    {
      "id": "terminal-completed",
      "kind": "terminal",
      "status": "completed"
    },
    {
      "id": "transcript-complete",
      "kind": "transcript_complete"
    },
    {
      "id": "mock-only",
      "kind": "network",
      "realPaperclipRequests": 0
    }
  ],
  "description": "The agent reads the current document revision before making an optimistic update.",
  "fixture": "../fixtures/companies/orientation.json",
  "id": "update-task-document",
  "prompt": "Read the current eval-notes document, then update it by appending a second line: Next assertion: context before action. Preserve the existing text and use the current revision id as baseRevisionId.",
  "schema": "paperclip-runner/eval-case/v1",
  "title": "Agent safely updates a task document"
}