Design gated sub-loop workflows with machine-verifiable definitions of done.
npx skills add https://github.com/VincentJiang06/skills --skill loop-constructor
Please help me install this Agent Skill. Skill: loop-constructor Source: https://github.com/VincentJiang06/skills/tree/main/skills/loop-constructor Command: npx skills add https://github.com/VincentJiang06/skills --skill loop-constructor
Medium and large AI-coding tasks often fail because they lack structured verification, clear role separation, and machine-checkable completion criteria. Without these, agents loop endlessly, produce slop, or require constant human intervention. This skill solves that by providing a rigorous selection procedure and deterministic linter to design closed-loop workflows that run autonomously.
Use the loop-constructor skill to design a gated sub-loop runbook for your medium or large autonomous coding task, emitted as a .loop/ runbook with machine-verifiable checks and an adversarial evaluator role.
High-intent search queries and answers about installing and using this skill.
Autonomous agent loops prevent endless looping and slop by decomposing coding tasks into gated sub-loops with machine-verifiable definitions of done. A deterministic linter validates the loop-design JSON to ensure clear role separation and deterministic completion criteria before rendering a runbook.
Autonomous AI agents loop endlessly or produce slop when tasks lack structured verification, clear role separation, and machine-checkable completion criteria. Applying a gated sub-loop design with adversarial evaluator roles and deterministic validation prevents these failures during autonomous execution.
Create machine-verifiable definitions of done by applying a rigorous D0-D6 decision tree to derive loop shape from runnable checks. This staged loop design process uses contract negotiation and adversarial evaluator roles to establish deterministic completion criteria validated by a linter.
The best way to structure semi-autonomous coding tasks is decomposing them into gated sub-loops with machine-verifiable definitions of done. This approach uses a D0-D6 selection procedure covering decomposition, pattern selection, autonomy level, parallelism, guards, and iteration profile for deterministic validation.
Yes, you can use a deterministic linter to validate autonomous workflow loop designs. The linter rejects hollow checks, missing contracts, or cyclic dependencies in the loop-design JSON, ensuring only passing designs with proper adversarial verification are rendered into a .loop/ runbook.
You need a gated sub-loop runbook for agent orchestration scenarios requiring autonomous execution, adversarial verification, and deterministic validation of completion criteria. It is designed for medium and large semi-autonomous AI-coding tasks that fail without structured verification and clear role separation.
fcircle-bizGenerate placeholder JPG images with embedded AI prompts and metadata.
Canhada-LabsDefine machine-verifiable goals and incorporate human review for agent loops.
SollanSystemsRoutes agent-loop objectives into design, contract, run, repair, eval, or flywheel workflows.
rohitg00Decompose complex tasks into atomic units with done criteria and dependency mapping.
