Suggested test: copilot/lazy-skill-blocks

[github-actions]

Motivation

Link to the gh-aw PR: github/gh-aw#46969 — feat(copilot-opt): reduce first-turn context size with lazy skill blocks and pre-agent guards

This change introduced lazy-load ## skill: blocks in workflow source files that are stripped from the initial system message and only loaded on demand. The existing test-copilot-skills-frontmatter.md only tests skills declared via the skills: frontmatter field, not inline ## skill: prompt blocks.

Proposed test

• Workflow file: test-copilot-inline-skill-block.md
• Trigger: workflow_dispatch
• Engine: copilot
• Safe output: create-issue
• Variant: standard

Minimal test prompt sketch

Define a workflow with an inline ## skill: my-skill block containing a small instruction set. The main prompt asks the agent to invoke the skill by name and confirm it received the skill content. The agent creates an issue whose body confirms the skill block was accessible at runtime (not stripped).

New fixtures or secrets needed

None — inline skill blocks are defined in the workflow source file itself, requiring no external fixture or secret.

Notes

The test-copilot-skills-frontmatter.md tests externally-referenced skills (GitHub-hosted). This would be the first test for inline skill block parsing and lazy-load behavior. Overlaps conceptually with test-copilot-inline-sub-agents.md but exercises a different feature (skill lazy-load vs. sub-agent dispatch).

Warning

Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• awmgmcpg

To allow these domains, add them to the network.allowed list in your workflow frontmatter:

network:
  allowed:
    - defaults
    - "awmgmcpg"

See Network Configuration for more information.

Generated by 🔍 Suggest New E2E Tests · 27.7 AIC · ⌖ 5.71 AIC · ⊞ 5.7K · ◷