paperclip-runner evalsstatic evidence

← Overview

Agent discovers an authorized optional capability

lazy-discover-agents

Lazy exposure requires discovery before scoped actor lookup.

Prompt

Use capability discovery to find the appropriate Paperclip operation, then list the agents in this company.

Fixture

../fixtures/companies/orientation.json

Authority

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

Current results

ModelConfigurationResultEstimated costRun time
claude-sonnet-5live-acpx-0.13.1-claude-acp-0.70.0-sonnet-5INFRAunknown2026-09-06T20:52:26.592Z
claude-sonnet-5live-claude-managed-agents-2026-04-01-sonnet-5PASS$0.019025 estimated2026-09-06T20:53:48.227Z
global.anthropic.claude-sonnet-4-6live-aws-agentcore-global-anthropic-claude-sonnet-4-6INFRAunknown2026-09-06T20:53:16.424Z
gpt-5.4-minilive-codex-pinnedINFRAunknownunknown
gpt-5.6-sollive-acpx-0.13.1-codex-acp-1.6.2-controlNot run——
openrouter/anthropic/claude-sonnet-5live-opencode-1.18.17-openrouter-claude-sonnet-5PASS$0.018901 estimated2026-09-06T20:55:25.010Z
openrouter/deepseek/deepseek-v4-flash-0731live-opencode-1.18.17-openrouter-deepseek-v4-flash-0731PASS$0.000729 estimated2026-09-06T20:58:30.018Z
openrouter/google/gemini-3.8-flashlive-opencode-1.18.17-openrouter-gemini-3.8-flashPASS$0.022734 estimated2026-09-06T20:55:41.886Z
openrouter/openai/gpt-6-astralive-opencode-1.18.17-openrouter-gpt-6-astraPASS$0.033213 estimated2026-09-06T20:57:08.422Z
openrouter/qwen/qwen3.8-max-0902live-opencode-1.18.17-openrouter-qwen3.8-max-0902PASS$0.010846 estimated2026-09-06T20:57:22.480Z
openrouter/z-ai/glm-5.3live-opencode-1.18.17-openrouter-glm-5.3PASS$0.007680 estimated2026-09-06T20:55:59.489Z

Assertions

discovery-loaded
evidence_match
{
  "evidenceKind": "tool_discovery",
  "id": "discovery-loaded",
  "kind": "evidence_match",
  "where": {
    "action": "loaded",
    "operationIds": {
      "includes": [
        "list_agents"
      ]
    }
  }
}
list-agents
semantic_operation
{
  "id": "list-agents",
  "kind": "semantic_operation",
  "maxCalls": 1,
  "minCalls": 1,
  "operationId": "list_agents",
  "requireSuccess": true
}
terminal-completed
terminal
{
  "id": "terminal-completed",
  "kind": "terminal",
  "status": "completed"
}
mock-only
network
{
  "id": "mock-only",
  "kind": "network",
  "realPaperclipRequests": 0
}
Full authored test definition
{
  "authority": {
    "actorId": "actor-engineer",
    "capabilities": [
      "discovery:agents:read"
    ],
    "explicitClaims": [
      "discovery:agents:read"
    ],
    "taskId": "task-orientation"
  },
  "checks": [
    {
      "evidenceKind": "tool_discovery",
      "id": "discovery-loaded",
      "kind": "evidence_match",
      "where": {
        "action": "loaded",
        "operationIds": {
          "includes": [
            "list_agents"
          ]
        }
      }
    },
    {
      "id": "list-agents",
      "kind": "semantic_operation",
      "maxCalls": 1,
      "minCalls": 1,
      "operationId": "list_agents",
      "requireSuccess": true
    },
    {
      "id": "terminal-completed",
      "kind": "terminal",
      "status": "completed"
    },
    {
      "id": "mock-only",
      "kind": "network",
      "realPaperclipRequests": 0
    }
  ],
  "description": "Lazy exposure requires discovery before scoped actor lookup.",
  "fixture": "../fixtures/companies/orientation.json",
  "id": "lazy-discover-agents",
  "prompt": "Use capability discovery to find the appropriate Paperclip operation, then list the agents in this company.",
  "schema": "paperclip-runner/eval-case/v1",
  "title": "Agent discovers an authorized optional capability",
  "toolExposure": "lazy"
}