Prompt control

Prompt and context rules for IDE coding agents. Shape what enters model context and what agents are instructed to do—redact secrets, block risky system prompts, keep org rules sticky.

Why it matters

Jailbreak-style overrides and accidental pastes of API keys into context bypass shell and path rules. Prompt control is the channel that governs instructions and context expansion before the model turn.

Allow and deny examples

Allow

• Repo conventions, approved skill prompts

Deny

• API keys in context, jailbreak-style overrides

How it works

When an agent expands prompts or context, Relay evaluates prompt rules before the model turn proceeds with that material. Denied expansions return a structured reason and are recorded for audit.

Harness prompt management versions org templates for production agents; Relay prompt control enforces IDE-session rules for Cursor and Claude.