Orchestrate AI coding agents through phase-locked workflows with security gates.
npx skills add https://github.com/Josemalyson/hes --skill harness-engineer-josemalyson
Please help me install this Agent Skill. Skill: harness-engineer Source: https://github.com/Josemalyson/hes/tree/main Command: npx skills add https://github.com/Josemalyson/hes --skill harness-engineer-josemalyson
This Skill provides a formal, scalable framework to orchestrate AI coding agents through a phase-locked lifecycle, ensuring repeatable workflows, traceability, and continual improvement.
Invoke /hes to bootstrap and begin the DISCOVERY phase; the harness will autonomously navigate gates, record outcomes, and prepare for the next feature.
High-intent search queries and answers about installing and using this skill.
You can orchestrate AI agents through a phase-locked workflow by applying a strict harness that enforces sequential phases like discovery, spec, design, and testing with security gates. This ensures repeatable execution and traceability.
Event sourcing for AI agents involves recording execution outcomes and decisions in an events.log file. This creates a persistent memory loop that enables context consolidation and continual improvement across future projects.
Yes, a state machine can enforce security gates by phase-locking the workflow. It ensures autonomous AI agents cannot advance from development phases to review without passing explicit security checks and harness-health validations.
To bootstrap a multi-agent project, you invoke the harness command to autonomously initialize the discovery phase. The harness then navigates gates, records outcomes to logs, and prepares for the next feature.
A learning loop improves orchestration by capturing lessons in a markdown file and logging events during execution. This consolidates context and feeds continual improvement back into the protocol-driven workflow.
Happyverse-TeamManage Ralph orchestration loops, monitor statuses, merge completed work, troubleshoot failures via CLI commands.
mejangoParse SARIF 2.1.0 files to extract and deduplicate findings.
Imad-OuteAutomate code reviews for correctness, security, and testability in SDLC phase 6.
wmasfoeOrchestrate parallel execution of independent tasks with lightweight evidence and acceptance criteria.
