meta-harness
CommunityManage scoped agent work with upstream-first fixes.
Software Engineering#session management#orchestration#agentic coding#OpenCode#scope decomposition#upstream-first repair
Authorthistleknot
Version1.0.0
Installs0
System Documentation
What problem does it solve?
The meta-harness prevents agent drift by giving the requester a manager layer that cleanly frames objectives, partitions scope, and delegates execution without mixing “what to do” with “how to do it.”
Core Features & Use Cases
- Objective framing & scope partitioning: Defines objective, constraints, and success criteria before dispatching any work.
- Parent/child session orchestration: Runs one scoped objective per session, resumes valid sessions, and forks parallel child sessions for independent issues.
- Upstream-first repair routing: Identifies upstream source surface → transformation/generation layer → downstream artifact, then fixes at the highest valid upstream layer and regenerates outputs.
- Operational handoff to agentic execution: Delegates a scoped task packet (expected outputs, validation commands, escalation conditions) to agentic-harness for concrete implementation.
- OpenCode session pattern support: Uses one-shot sessions, resuming by session id, and forking per task for parallel work.
Quick Start
Ask your AI to run in manager mode using OpenCode, and explicitly provide the objective, constraints, and success criteria before it delegates each scoped issue to agentic-harness.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: meta-harness Download link: https://github.com/thistleknot/skills/archive/main.zip#meta-harness Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 471,000+ vetted skills library on demand.