Automate adversarial multi-agent validation and iterative correction of artifacts.
npx skills add https://github.com/iamsteveholmes/momentum --skill avfl
Please help me install this Agent Skill. Skill: avfl Source: https://github.com/iamsteveholmes/momentum/tree/main/skills/momentum/skills/avfl Command: npx skills add https://github.com/iamsteveholmes/momentum --skill avfl
AVFL automates adversarial, multi-agent validation to improve artifact quality and reduce hallucinations by coordinating dual reviewers, corpus cross-document checks, and a fixed workflow.
Launch AVFL through momentum with the default full workflow to validate and fix an artifact.
High-intent search queries and answers about installing and using this skill.
Adversarial multi-agent validation uses dual reviewers, an Enumerator and an Adversary, to assess artifacts across structural, accuracy, coherence, and domain lenses, enforcing evidence-backed findings to improve quality and reduce hallucinations.
You automate iterative correction by running a fixed workflow that validates artifacts, consolidates findings with strict scoring, and applies fixes, looping through up to four iterations until quality thresholds are met.
Yes, you can use corpus-mode validation to perform cross-document consistency and corpus completeness checks when validating multiple files, ensuring broader coverage across related artifacts.
Yes, the framework enforces evidence-backed findings with strict scoring and logging, requiring reviewers to substantiate their assessments across all four validation lenses.
The main limitation is that the iterative correction loop runs a maximum of four iterations, meaning artifacts requiring extensive fixes beyond this threshold may need manual intervention or additional runs.
jscraikExecute authorized interrogation workflows for macOS apps, web apps, and repositories.
michaelarutyunovExport simulated interview artifacts into a timestamped folder under reports/interviews.
kenkenmainCoordinate a six-phase A0-A5 swarm pipeline with agent teams and dual-channel communication.
nguyenphpRun linters, type checkers, and security audits across Node.js and Python projects.
