Back to Agno

Test Log: 23_skills

cookbook/05_agent_os/23_skills/TEST_LOG.md

2.8.22.3 KB
Original Source

Test Log: 23_skills

Tested on 2026-07-24 against Agno source commit 45bfff9f2aa6ec11b7386c3cd3bf6d1141d005dc.

The lesson was loaded from the rewrite worktree, its scripts were executed directly, and its Agent was exercised through a real model-backed AgentOS run.

basic.py

Status: PASS

Test mode: LIVE

Description: Started the standalone AgentOS on port 8893, checked health and discovery, then asked gpt-5.5 to load and execute the local system-info skill through the non-streaming Agent run API.

Result: /health returned ok; /config returned skills-agent-os and skills-agent. Run 9c03f3c9-5ace-4e57-8844-7ac24f49969a completed after recorded calls to get_skill_instructions and get_skill_script with execute=True. The script result had return code 0, empty standard error, and JSON standard output for host Ios-MacBook-Pro.local, OS Darwin, and Python 3.14.5.


sample_skills/system-info/scripts/get_system_info.py

Status: PASS

Test mode: LIVE

Description: Ran the script directly with the demo Python environment and again through the executable skill tool.

Result: Direct execution exited 0 and printed one valid JSON document containing architecture, UTC time, hostname, OS release, processor, Python executable, and Python version. Skill-tool execution also exited 0; its standard output was the JSON consumed by the live AgentOS run.


sample_skills/system-info/scripts/list_directory.py

Status: PASS

Test mode: LIVE

Description: Ran the script directly against the sample skill directory and through the executable skill tool with args=["."].

Result: Both executions exited 0 and printed valid JSON. The direct run reported exactly the scripts/ directory and SKILL.md, sorted with the directory first.


Validation

  • Both direct scripts exited 0 and produced parseable JSON.
  • LocalSkills validated and loaded system-info; Skills registered get_skill_instructions, get_skill_reference, and get_skill_script.
  • The focused loader and skill-tool suites passed all 71 tests.
  • Recursive pattern validation checked exactly 3 Python files with 0 violations.
  • Targeted Ruff format and check, Python compilation, inventory parity, local Markdown links, legacy deletion, and git diff --check passed.