Skip to content
Get packs

Honeycomb Tracing

Trace Paperclip runs and CI suites in Honeycomb and explain where time goes

Source Download skill
Repository
paperclipai/paperclip
Stars
110,078
Licence
MIT
Files
5 files

What it does

This skill teaches an AI agent how to investigate slow or failing Paperclip runs, requests, and CI suites using Honeycomb trace data. Once installed, the agent knows how to find the right trace from a run ID, read the timeline of spans (timed steps), separate serial from parallel time, and suggest the smallest useful next experiment. It also covers configuring opt-in OpenTelemetry and adding missing spans, but it explicitly says tracing alone does not authorize replaying tasks, changing credentials, or deploying instrumentation.

Great for

  • Engineers maintaining the Paperclip project who need to diagnose latency in agent runs
  • Developers investigating slow CI suites and which test files delay them
  • Operators with Honeycomb access who want a structured way to read Paperclip traces

Skip it if

  • People without access to a Paperclip instance or Honeycomb account, since the workflow depends on real run data
  • Anyone expecting a general Honeycomb tutorial; it is specific to Paperclip's span names and repository layout
  • Users who might paste API keys into the agent, since the skill warns secrets must stay out of commands and reports but the risk remains if handled carelessly

Try asking

  • “Find the Honeycomb trace for run UUID 1234abcd and explain which intervals dominated its elapsed time”
  • “Why does this CI suite take 40 minutes? Profile which test files delay the whole run”
  • “Design a readable waterfall for the issue page load so a teammate can understand it”
  • “Add a span around workspace finalization and tell me where it should go”
Step-by-step guide: how to use Honeycomb TracingInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 5 original files, the MIT licence and an attribution note.

    honeycomb-tracing.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip honeycomb-tracing.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip honeycomb-tracing.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent the exact run UUID and a UTC time window; a task identifier is not enough
  • Ask for diagnosis only if you do not want the agent to replay tasks or change configuration
  • Verify the Honeycomb team, environment, and dataset before comparing data, since the defaults may not match your deployment

What's inside

  • agents/openai.yaml
  • references/ci.md
  • references/runtime.md
  • references/trace-design.md
  • SKILL.md

Redistributed unmodified from paperclipai/paperclip under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.