skills.rest
ax-llm
ax-llmOFFICIAL
Explore all skills by this author →
GitHub

ax-agent-observability

Add structured observability callbacks and accessors to AxAgent runs.

2.9k|186|Updated Feb 23, 2023
Search Tags:#observability#tracing#telemetry#callbacks#axagent#usage accounting
One-click install
npx skills add https://github.com/ax-llm/ax --skill ax-agent-observability
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ax-agent-observability
Source: https://github.com/ax-llm/ax/tree/main/website/static/typescript/.well-known/agent-skills/ax-agent-observability
Command: npx skills add https://github.com/ax-llm/ax --skill ax-agent-observability

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you add the right observability hooks to AxAgent so you can see what happened during a run without scraping prompts or logs by hand.

Core Features & Use Cases

  • Structured turn telemetry with actorTurnCallback for code, results, thoughts, stage, usage, and chat-log messages.
  • Context-pressure and compaction events with onContextEvent for checkpoints, tombstones, relevance ranking, and field promotion.
  • Live progress updates with agentStatusCallback, plus function-call tracing and usage or trace retrieval through onFunctionCall, getChatLog, getUsage, getStagedUsage, getTraces, and resetUsage.
  • Use it when debugging agent behavior, building dashboards, measuring prompt pressure, or auditing execution flow in production.

Quick Start

Use the ax-agent-observability skill to add the smallest observability hook that exposes the exact AxAgent telemetry you need.

Frequently Asked Questions about ax-agent-observability

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add observability and tracing to AxAgent runs?▼

Add structured observability to AxAgent runs by attaching callbacks like actorTurnCallback for turn telemetry and onFunctionCall for runtime function-call tracing. This exposes execution flow without scraping prompts manually.

What is the best way to monitor token usage and context pressure in LLM agents?▼

Monitor token usage and context pressure using getUsage, getStagedUsage, and onContextEvent. These accessors expose compaction checkpoints, relevance ranking, and field promotion events to measure prompt pressure during execution.

How do I get real-time progress updates and chat logs during agent execution?▼

Get real-time progress updates and inspect chat logs by implementing agentStatusCallback for live status reporting, and retrieve detailed conversation history using the getChatLog accessor to audit execution flow.

Can I debug AxAgent function calls and collect program traces in production?▼

Debug AxAgent function calls and collect program traces in production using onFunctionCall for tracing and getTraces for retrieval. These hooks capture runtime execution details for auditing without manual log scraping.

When should I use resetUsage instead of getUsage for token accounting?▼

Use resetUsage to clear accumulated token counts and start a new accounting period, whereas getUsage retrieves the current cumulative totals. Reset is for isolating usage metrics per run or stage.

Does ax-agent-observability support building dashboards for agent behavior?▼

It supports building dashboards by providing structured telemetry through callbacks like actorTurnCallback and agentStatusCallback, which expose code, results, thoughts, stage, and usage data suitable for dashboard visualization.

Related Skills

Explore More Software Engineering Skills→
ax-llmax-llm
2.9k

ax-go-agent-observability

Inspect Ax Go agent traces, usage, runtime envelopes, action logs, and replay state.

Official
Advanced
elasticelastic
546

elasticsearch-audit

Configure and query Elasticsearch security audit logs via cluster settings API.

Official
Intermediate
ax-llmax-llm
2.9k

ax-flow

Generate typed AxFlow workflow code for AI pipelines.

Official
Advanced
berea-chberea-ch

customize-sdk-hooks

Implement custom code hooks in Speakeasy-generated SDKs for lifecycle events.

Official
Advanced