YES.md 日本語版v1.1.0

A developer is debugging a production API failure. The AI agent must avoid assumptions like 'probably a network issue' and instead run curl commands to gather evidence, check logs, and validate fixes with tests before reporting completion, ensuring no related bugs are overlooked.

An IT administrator needs to update a critical config.yaml file. The AI agent must first create a backup, grep for dependencies, verify no file locks with lsof, and test the change by restarting services and checking logs to prevent downtime.

During a database migration, errors occur. The AI agent must not guess causes but use Bash to check permissions, read logs, and search documentation. After applying a fix, it runs queries to confirm data integrity and checks for edge cases like null values.

A team is deploying a new containerized application. The AI agent ensures safety by verifying container health with docker ps, checking for uncommitted changes, and testing endpoints with curl post-deployment, avoiding manual user checks.

Integrating a third-party API fails intermittently. The AI agent avoids surface fixes by using WebSearch for error messages, reading API docs with Read, and creating a minimal reproduction case after multiple failures to isolate the root cause.

Offer this skill as part of a premium support tier for SaaS platforms, reducing customer tickets by ensuring AI agents provide accurate, evidence-based solutions without manual escalation, improving efficiency.

Integrate YES.md into DevOps tools like CI/CD pipelines or monitoring systems, where it enforces safety gates and evidence rules during deployments and debugging, sold as an add-on feature.

Provide consulting services to train AI development teams on implementing YES.md principles, including workshops on debugging protocols and safety checks, tailored for industries like fintech or healthcare.