Why Agent Forge

This column states claims and constraints; mechanism detail lives in the governance and architecture columns. Each article addresses one premise that is often skipped, and links back to implementation pages you can check.

Articles in this column

• Coding agents do not run inside a platform—how do you govern them?
• Data must stay in place—how do you add governance for agents?
• Why having logs is not the same as accountability
• Privilege escalation is rarely fixed by adding an approval console
• Why private deployments need modifiable source

Claims mapped to mechanism pages

Article	Claim	Mechanism pages
Cross-process governance	The gate sits outside the executing process	HITL · why-cs
Data not moving	Governance deploys next to the data	Sandbox · Deployment
Verifiable audit	Queryable logs ≠ accountability	Hash chain
Permission isolation	Model and binding first; approval UI second	RBAC · Identity
Source modifiable	Auditable and patchable is a private-deploy baseline	Seam · Baseline