Skip to main content
Use skills to add instructions or context to your Canvas investigations. Skills help Canvas remember important details about your applications, your team, or preferred ways of working. Create a new skill whenever you find yourself repeating the same steps in a Canvas investigation or want to share a particular multi-step investigation workflow with your team. Canvas skills follow the Agent Skills standard. Canvas only uses skills during relevant investigations or you can invoke skills
by typing /skill-name in the chat. You can access your team’s Canvas skills in two ways:
  • Navigate to Canvas > Skills and Knowledge
  • Enter forward slash (/) in a Canvas chat and select Manage Skills

​
Creating skills

To create a new skill, navigate to Canvas > Skills and Knowledge and select New Skill in the top right.
1

Give your skill a name

  • 1-64 characters
  • Can only use lowercase alphanumeric characters (a-z, 0-9) and hyphens (-)
  • Cannot start or end with a hyphen (-)
2

Add a description

Describe what your skill does and when the Canvas agent should use it.You can include trigger phrases in your skill’s description to help Canvas better understand when to run your skill.
Trigger phrases: "which users are getting rate limited",
"debug errors at checkout","who is using my feature","traffic trends"
3

Add your instructions and context

Write instructions for the Canvas agent to follow when your skill activates.Skill content can be provided as plain text or Markdown.
# Investigate API Latency

When this skill is activated:
- Identify affected services
- Check recent deploys
- Compare current latency to baseline
- Suggest likely root causes

Be concise and focus on actionable insights.

​
Editing and managing skills

You can access your team’s Canvas skills in two ways:
  • Navigate to Canvas > Skills and Knowledge
  • Enter a forward slash (/) in the chat and select Manage Skills
From the Skills and Knowledge page you can:
  • Create a new skill by selecting New Skill.
  • Edit skills using the edit button.
  • Delete skills using the delete (trash can) button.

​
Built-in skills

Canvas includes the following skills by default. These skills cannot be removed or modified.
  • conversation-investigation: Debug AI sessions, conversations, or AI workflows captured via OpenTelemetry GenAI semantic conventions using Honeycomb’s Agent Timeline.
  • metrics-queries: How to query metrics datasets in Honeycomb.
  • observability-fundamentals: Answers conceptual questions about observability and gives tool-agnostic guidance.
  • production-investigation: Debug production issues using the core analysis loop and Honeycomb features.
  • query-patterns: How to write effective Honeycomb queries.
  • slos-and-triggers: How to design SLOs, when to use SLOs or triggers, and how to interpret an SLO status.
  • summarize: Summarize a Canvas so you can quickly get up to speed.