Skip to content

:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.

990 B
Raw Blame
import { HumanizedActivityLogItem } from './humanizeActivity'

// Matches AGENT_TRIGGER_JOB_TYPE in posthog/models/activity_logging/activity_log.py. Product
// triggers carry job ids that are not sandbox tasks, so only this job type links to a task.
const AGENT_TRIGGER_JOB_TYPE = 'agent'

export interface AgentAttribution {
    /** Self-reported by the agent, never verified. */
    intent: string | null
    /** Bound to the agent's token by the server. Intent alone cannot verify a task. */
    taskId: string | null
}

export function parseAgentAttribution(logItem: HumanizedActivityLogItem): AgentAttribution | null {
    const trigger = logItem.unprocessed?.detail?.trigger
    if (trigger?.job_type !== AGENT_TRIGGER_JOB_TYPE) {
        return null
    }

    const intent = typeof trigger.payload?.intent === 'string' ? trigger.payload.intent : null
    const taskId = trigger.job_id || null
    if (!intent && !taskId) {
        return null
    }

    return { intent, taskId }
}