Implement 15 rules to prevent fabrication, constraint violations, and incomplete work.
npx skills add https://github.com/av/harbor --skill discipline-av
Please help me install this Agent Skill. Skill: discipline Source: https://github.com/av/harbor/tree/main/.agents/skills/discipline Command: npx skills add https://github.com/av/harbor --skill discipline-av
This Skill ensures that common agent failure modes are prevented through a robust set of rules, enhancing efficiency and reducing errors.
Load the 'discipline' skill to prevent agent failures in your workflow.
High-intent search queries and answers about installing and using this skill.
Load an agent operating protocol before starting tasks to prevent failures like fabrication and incomplete work. This protocol applies 15 rules covering constraint tracking, verification, scoping, and communication to enforce best practices and enhance reliability.
An agent operating protocol for failure prevention is a robust set of 15 behavioral rules ensuring agents do not ignore user constraints or fabricate data. It standardizes best practices to enhance efficiency and reduce errors in operational environments.
Enforce constraint tracking and verification in agent management by applying a structured operating protocol with 15 dedicated rules. This approach standardizes retry discipline and communication to prevent common failure modes and reduce rework.
This agent operating protocol is designed specifically for Ivan's environment to prevent agent failures. It focuses on best practices for agent behavior within that context to enhance reliability and efficiency.
AI agents ignore user constraints and deliver incomplete work due to a lack of operational discipline. Implementing a protocol with 15 failure prevention rules covering scoping, verification, and communication corrects these common failure modes.
The best way to enhance agent reliability and efficiency is to implement a comprehensive operating protocol. This prevents common failure modes through 15 rules covering fabrication, retry discipline, and communication, reducing the need for rework.
KentoShimizuDefine measurable quality constraints for approved functional requirements baselines.
hiddink-aiScore user requests for ambiguity across scope, technical, and acceptance factors.
beyonaiStandardize agent tool development through multi-phase audits and cross-modal evaluations.
bugrabilgeDesign benchmark test suites for evaluating LLM agent reliability.
