Minecraft-Console-Client/.skills/mcc-dev-workflow/evals/evals.json

41 lines
2.3 KiB
JSON

{
"skill_name": "mcc-dev-workflow",
"evals": [
{
"id": 1,
"prompt": "Build MCC, start the local 1.21.11 vanilla server from an MCC_SERVERS root, connect MCC with a temporary config, and verify a successful join plus one inventory command.",
"expected_output": "The workflow uses a real local server, a temp MCC config, and concrete log evidence for both the join and the MCC command.",
"files": [],
"expectations": [
"The workflow uses a real local 1.21.11 server instead of only reading code.",
"The workflow uses MCC_SERVERS-aware tooling or documents the server root explicitly.",
"The workflow uses a temporary MCC config instead of relying on the repo-root config.",
"The result includes join evidence from MCC output and the server log."
]
},
{
"id": 2,
"prompt": "Debug a flaky local MCC startup on 1.21.11 by checking for stale tmux server sessions, waiting for server readiness, and retrying early RCON commands before blaming protocol code.",
"expected_output": "The response treats tmux sessions as shared state, distinguishes server startup from RCON readiness, and uses the real local workflow rather than pure speculation.",
"files": [],
"expectations": [
"The workflow checks or mentions stale mc-* tmux sessions.",
"The workflow distinguishes Done from RCON readiness.",
"The workflow retries or advises retrying early RCON commands.",
"The workflow keeps the debugging loop grounded in real local commands."
]
},
{
"id": 3,
"prompt": "Run a repeatable local MCC debug loop for 1.21.11 that compiles the client, connects with movement, inventory, and entity handling enabled, and leaves enough evidence to inspect a regression afterward.",
"expected_output": "The response follows a real build-run-test-inspect loop and captures enough log evidence to support follow-up debugging.",
"files": [],
"expectations": [
"The workflow builds MCC before the run.",
"The workflow enables terrain, inventory, and entity handling for the scripted run.",
"The workflow captures or points to concrete log locations.",
"The workflow prefers a temp config for repeatability."
]
}
]
}