Verify AI sub-agent outputs through deterministic gates and council review.
npx skills add https://github.com/Nefas11/governed-agents --skill governed-agents
Please help me install this Agent Skill. Skill: governed-agents Source: https://github.com/Nefas11/governed-agents/tree/main Command: npx skills add https://github.com/Nefas11/governed-agents --skill governed-agents
Consolidates verification and reputation management to prevent hallucinated success by AI sub-agents. It organizes a deterministic 4-gate coding verification suite and a 3-layer verification pipeline for open-ended tasks.
Spawn a governed sub-agent for a coding task or open-ended task and observe automatic gating, council verification, and reputation updates.
High-intent search queries and answers about installing and using this skill.
Verify AI sub-agent coding outputs using deterministic 4-gate verification checks for files, tests, lint, and AST. This ensures deliverables exist and meet quality checks before integration.
Govern open-ended AI tasks using a 3-layer verification pipeline: Structural Gate, Grounding Gate, and an LLM Council. This prevents hallucinated success by validating structural integrity and grounding.
Reputation tracking uses an EMA-based scoring system and a persistent ledger to monitor AI agent performance. It updates reputations across tasks for progressive supervision and risk-aware task allocation.
Configurable task profiles support scalable AI governance by tailoring verification rules to specific task types. This allows progressive supervision and risk-aware task allocation across diverse workflows.
AI sub-agents hallucinate successful task completion due to unreliable outputs lacking deterministic verification. Applying a 4-gate pipeline and 3-layer council verification prevents this by validating results.
ArvorCoCreate or update a project constitution defining coding principles and governance rules.
DaddyElonMusk69Generate macro factor signals for cross-asset allocation from economic data.
chroniicallydiistractedRun CI checks, categorize lint/type/test errors, and apply fixes.
as4584Evaluate LLM applications with automated metrics and human feedback.
