jeong-sik

sangokushi-2-end-command

"One call to end the current province's commands in Sangokushi II (0, Return, y): presses the verified sequence, waits for the settled screen, returns the observation. Use only at the province (0-19)? prompt."

jeong-sik 3 Updated 1mo ago
GitHub

Install

npx skillscat add jeong-sik/masc/sangokushi-2-end-command

Install via the SkillsCat registry.

SKILL.md

Sangokushi II: end the province's commands

The verified province-command end for Sangokushi II, observed as the province
10 to 18 transition. One call presses 0, Return, y in sequence, advances
until the screen settles (masc_msx_step_until_change), and returns the
settled observation with its frame and artifact.

Use it only when the visible prompt is the province command prompt of the
province you are finishing; a macro pressed on another screen is just keys.
The game knowledge this macro comes from — menus, save flow, media pitfalls —
lives in the sangokushi-2 Skill. The machine is shared: an unexpected
result may mean another driver intervened between the presses and the settle;
re-observe before retrying (the msx-play Skill, shared machine).

[[compositions]]
name = "sangokushi-2-end-command"
description = "End the current province's commands (0, Return, y) and return the settled screen."
execution = "inline"

[[compositions.nodes]]
id = "press"
tool = "masc_msx_press"
[compositions.nodes.input]
kind = "object"

[[compositions.nodes.input.fields]]
name = "keys"
[compositions.nodes.input.fields.value]
kind = "literal"
value = ["0", "Return", "y"]

[[compositions.nodes.input.fields]]
name = "sequence"
[compositions.nodes.input.fields.value]
kind = "literal"
value = true

[[compositions.nodes]]
id = "settle"
tool = "masc_msx_step_until_change"
after = ["press"]
[compositions.nodes.input]
kind = "literal"
value = {}

[[compositions.nodes]]
id = "screen"
tool = "masc_msx_screen"
after = ["settle"]
[compositions.nodes.input]
kind = "literal"
value = {}