Pstack Model Configurator
Configures role-based model allocations and reasoning budgets for pstack AI coding subagents.
Configures and optimizes model assignments and reasoning effort budgets across specialized AI subagent roles within the pstack framework. It automatically detects accessible model slugs, prompts developers to select budget profiles ranging from small to unlimited reasoning effort, and writes persistent Cursor rules to govern multi-agent task execution, bug fixing, reviews, and architectural planning.
主な機能
01Automated detection and validation of accessible AI model slugs
02Granular role-to-model mapping across coding, reviewing, and judge tasks
03Parent model inheritance support via inherit-parent and auto aliases
04Tiered reasoning budget presets from small (medium effort) to unlimited (max effort)
05Generates persistent, idempotent Cursor rules via pstack-models.mdc
0646 GitHub stars
ユースケース
01Calibrating model reasoning budgets to balance API consumption against depth of analysis
02Assigning distinct LLMs to specialized roles such as architecture runners, judges, and code synthesizers
03Standardizing multi-agent workspace configurations across Cursor and Claude Code environments