Skip to main content
POST
/
agent
/
negotiation
Assess Task Capability
curl --request POST \
  --url http://localhost:3773/agent/negotiation \
  --header 'Content-Type: application/json' \
  --data '
{
  "task_summary": "Answer questions about machine learning concepts",
  "task_details": "User needs help understanding neural networks and deep learning",
  "input_mime_types": [
    "text/plain",
    "application/json"
  ],
  "output_mime_types": [
    "application/json"
  ],
  "max_latency_ms": 5000,
  "max_cost_amount": 0.01,
  "required_tools": [],
  "forbidden_tools": [],
  "min_score": 0.5,
  "weights": {
    "skill_match": 0.55,
    "io_compatibility": 0.2,
    "performance": 0.15,
    "load": 0.05,
    "cost": 0.05
  }
}
'
{
  "accepted": true,
  "confidence": 0.85,
  "scores": {
    "skill_match": 0.9,
    "io_compatibility": 1,
    "performance": 0.8,
    "load": 0.7,
    "cost": 0.9
  },
  "overall_score": 0.87,
  "reasoning": "Agent has strong capability in question answering with compatible IO formats"
}

Body

application/json
task_summary
string
required

Brief summary of the task

Maximum string length: 10000
task_details
string

Detailed task description

input_mime_types
string[]

Accepted input MIME types

output_mime_types
string[]

Accepted output MIME types

max_latency_ms
integer

Maximum acceptable latency in milliseconds

max_cost_amount
number

Maximum acceptable cost

required_tools
string[]

Tools that must be available

forbidden_tools
string[]

Tools that must not be used

min_score
number

Minimum acceptance score threshold

Required range: 0 <= x <= 1
weights
object

Scoring weights for different factors

Response

Capability assessment result

accepted
boolean
required

Whether the agent accepts the task

confidence
number
required

Confidence in the assessment

Required range: 0 <= x <= 1
overall_score
number
required

Weighted overall score

Required range: 0 <= x <= 1
scores
object

Individual scoring factors

reasoning
string

Explanation of the assessment