Enforce the Specify-Limit-Isolate-Create-Evaluate workflow for subagent orchestration.
npx skills add https://github.com/riez/dots --skill slice-agent-harness
Please help me install this Agent Skill. Skill: slice-agent-harness Source: https://github.com/riez/dots/tree/main/.config/agentic/skills/slice-agent-harness Command: npx skills add https://github.com/riez/dots --skill slice-agent-harness
The SLICE harness provides a mandatory, governance-driven framework for all agent orchestration. It enforces the SLICE methodology—Specify context, Limit tools, Isolate agents, Create task specs, and Evaluate quality—to ensure safe, auditable, and effective subagent coordination before dispatching any multi-step task.
Describe a task for orchestration by applying SLICE steps before dispatching any subagent.
High-intent search queries and answers about installing and using this skill.
The SLICE methodology for subagent orchestration is a governance-driven workflow that enforces Specify, Limit, Isolate, Create, and Evaluate steps. It ensures safe, auditable automation by applying mandatory checklists for context provisioning, restricted tooling, and isolated execution.
You can enforce output validation and task isolation in multi-agent systems by applying a structured orchestration harness. This framework mandates complete task specifications, standardized output formats, and restricted tooling to achieve auditable and isolated subagent execution.
To create task specs for autonomous agent workflows, apply a mandatory checklist that defines context provisioning, tool limitations, and measurable quality criteria. This ensures subagents receive complete, structured specifications before dispatch.
The best way to govern subagent execution and prevent misconfiguration is to use a structured orchestration harness. It defines orchestrator responsibilities, enforces standardized templates, and provides anti-pattern avoidance guidance to reduce risk.
You should not use an isolated agent execution framework for simple, single-step tasks that do not require careful context provisioning or measurable quality criteria. The framework is designed for complex, multi-step automation requiring strict governance and traceability.
wan-huiyanSpawn successor agents to maintain context continuity in long-running workflows.
slowman2084Convert platform test stdout logs into ShareGPT-format JSON files.
rozxVerify code changes against specs, tasks, and design artifacts.
oci-ai-architectsPlan multi-agent systems using the AGENT Blueprint architecture.
