Orchestrate research-run campaigns with metric-driven evaluation and auto-rollback.
npx skills add https://github.com/Borda/AI-Rig --skill run-borda
Please help me install this Agent Skill. Skill: run Source: https://github.com/Borda/AI-Rig/tree/main/plugins/research/skills/run Command: npx skills add https://github.com/Borda/AI-Rig --skill run-borda
Sustained metric-improvement loop with atomic commits, auto-rollback, and experiment logging. Iterates with specialist agents, commits atomically, auto-rolls back on regression. Accepts a program.md file path. Supports --resume, --team, --colab, --codex, --researcher, --architect, --journal, --hypothesis.
Provide a program.md and run the run command to start an automated, metric-driven campaign.
High-intent search queries and answers about installing and using this skill.
To automate an AI experiment campaign, you provide a program.md file and run the command to orchestrate ideation, sandboxing, and metric evaluation, automatically tracking iterations and recording metrics for continuous improvement.
Yes, you can run automated experiment campaigns on Docker, Colab, or local runtimes. The tool orchestrates metric evaluation and ideation within these sandbox environments to safely test program improvements.
Auto-rollback works by enforcing atomic commits during experimentation and automatically reverting changes when a metric regression is detected, ensuring your target program remains stable and safe.
The best way to log experiment metrics is through comprehensive JSONL metrics and diary entries, which record iterations and experiment logs to maintain full traceability of your metric-driven campaigns.
Yes, you need a program.md file to start the metric-improvement loop. It provides program-driven control alongside flags like --resume, --team, and --hypothesis to guide the campaign orchestration.
pingmepiAutomate PM-OS installation and configuration with pm_os_install.py.
reubenjohnConsolidate daily Diary entries into concise, pattern-rich memory notes.
tuanfungImplement trunk-based Git workflows with atomic commits and feature flags.
SjdjdiejdrirhdkjejOrchestrate staging validation with smoke tests, migration checks, and rollback testing.
