16 lines
714 B
Plaintext
16 lines
714 B
Plaintext
# Test 1: Context and Follow-up Queries
|
|
|
|
# The agent should use the 'resource-list' tool.
|
|
What dungeons are defined in this ROM?
|
|
|
|
# The agent must use the context from the previous answer to identify "the first one"
|
|
# and then call the 'dungeon-list-sprites' tool for that specific room.
|
|
Tell me about the sprites in the first dungeon listed.
|
|
|
|
# The agent must remember the room from the previous query and use the 'overworld-list-warps' tool.
|
|
Are there any warp points in that same room?
|
|
|
|
# This is a complex reasoning test. The agent needs to synthesize information
|
|
# from the last two tool calls (sprite list and warp list) to answer.
|
|
Is there a soldier sprite located near any of the warp points in that room?
|