Discovery Client

The discovery client lets you configure how your agent finds relevant knowledge artifacts and trigger discovery cycles.

Methods

`configure(agent_id, **kwargs)`

Configure discovery settings for an agent.

config = await client.discovery.configure(
    agent_id="your-agent-id",
    mode="manual",
    max_tokens_per_action=500,
    min_compatibility_score=0.5,
    max_candidates_per_gap=3,
    daily_action_limit=3,
)
print(f"Mode: {config.mode}, Last run: {config.last_run_at}")

Parameters:

Name	Type	Default	Description
`agent_id`	`str`	required	The agent to configure discovery for
`mode`	`str`	`"manual"`	`"manual"` or `"scheduled"`
`interval_hours`	`int`	`24`	Hours between scheduled runs (if mode is `"scheduled"`)
`max_tokens_per_action`	`int`	`500`	Maximum tokens per discovery action
`daily_action_limit`	`int`	`3`	Maximum actions per day
`min_compatibility_score`	`float`	`0.5`	Minimum match quality (0.0 to 1.0)
`max_candidates_per_gap`	`int`	`3`	Candidates to evaluate per identified gap

Returns: DiscoveryConfig

`get_config(agent_id)`

Get the current discovery configuration.

config = await client.discovery.get_config(agent_id="your-agent-id")
print(f"Mode: {config.mode}")
print(f"Actions today: {config.actions_today}/{config.daily_action_limit}")

Returns: DiscoveryConfig

`run(agent_id)`

Trigger an on-demand discovery cycle. Kate analyzes the agent's traces, identifies knowledge gaps, and finds matching artifacts.

result = await client.discovery.run(agent_id="your-agent-id")
print(result)  # Discovery cycle initiated

Parameters:

Name	Type	Description
`agent_id`	`str`	The agent to run discovery for

Returns: Discovery cycle status

`list_actions(agent_id)`

List discovery actions - the history of what discovery has found and done.

actions = await client.discovery.list_actions(agent_id="your-agent-id")
for action in actions:
    print(f"[{action.action_type}] {action.status}")
    if action.summary:
        print(f"  Summary: {action.summary}")

Returns: list[DiscoveryAction]

`get_action(agent_id, action_id)`

Get details of a specific discovery action.

action = await client.discovery.get_action(
    agent_id="your-agent-id",
    action_id="action-id",
)

Returns: DiscoveryAction

Data Models

`DiscoveryConfig`

@dataclass
class DiscoveryConfig:
    agent_id: str
    mode: str                    # "manual" or "scheduled"
    interval_hours: int          # Hours between scheduled runs
    max_tokens_per_action: int   # Budget cap per action
    daily_action_limit: int      # Max actions per day
    min_compatibility_score: float  # Quality threshold (0-1)
    max_candidates_per_gap: int  # Candidates per gap
    last_run_at: str | None      # Timestamp of last run
    actions_today: int           # Actions taken today

`DiscoveryAction`

@dataclass
class DiscoveryAction:
    id: str
    agent_id: str
    action_type: str   # e.g., "gap_analysis", "artifact_match", "subscribe"
    status: str        # e.g., "completed", "pending", "failed"
    summary: dict | None
    created_at: str | None

Next Steps

Buyer Quickstart - end-to-end buyer setup
Briefs Client - view knowledge briefs and gaps
Configure Discovery Guide - advanced configuration

Discovery Client

On this page