Policy

Policy controls what Murph does with a drafted response when a session is running. It is not a factuality checker.

Murph Agent is the preferred way to create or adjust a custom profile:

murph agent

Ask it to inspect your current policy, create a new profile, preview the result, and select it. Murph Agent can write policies/*.md without source-edit mode because policy files are part of the normal Plugin+Config authority surface.

Shipped profiles

Murph ships role-oriented profiles:

• default
• engineering
• product
• sales
• marketing
• leadership
• yolo

The role profiles are conservative and keep auto-send off. yolo is the maximum-autonomy preset for trusted local runs after you have verified setup and behavior. Runtime grounding still checks that required read/context work happens before answering.

Custom profile files

Profiles live in policies/*.md and use a metadata header plus body notes:

name: custom
description: My custom policy.
blockedTopics: payroll details, legal advice
alwaysQueueTopics: pricing, customer commitments
blockedActions:
allowAutoSend: no
requireGroundingForFacts: yes
preferAskWhenUncertain: yes
notes: keep replies concise, avoid promises
---
Extra instructions for Murph when this profile is active.

Use the CLI to inspect, preview, and select profiles:

murph policy profiles
murph policy preview --profile custom --mode auto_send_low_risk
murph policy set --profile custom

How policy runs

Murph keeps operational hard stops before the agent: no matching session, unknown target user, owner-authored events, expired sessions, and similar runtime conditions stop without drafting.

When those hard stops pass, the main agent drafts first. Then a small no-tool policy execution classifier reviews the request, policy, grounding status, and proposed action and returns send, queue, or abstain.

The deterministic final gate remains authoritative. dry_run, manual_review, allowAutoSend: no, blocked topics, blocked actions, high-risk skill context, unsupported action types, and low-confidence classifier sends can still force queue or abstain. Grounding is separate runtime behavior: it checks whether required read/context tools were attempted, but it does not prove factual correctness.

Session mode still matters

Policy profiles do not bypass session mode:

• manual_review queues actions for review.
• dry_run records decisions without side effects.
• auto_send_low_risk can auto-send only when the selected policy allows it.

Runtime hard stops still apply for empty context, out-of-scope threads, high-risk skill context, unsupported action types, and messages Murph cannot safely send.

YOLO profile

Use yolo when you intentionally want the least restrictive action profile:

murph policy set --profile yolo

yolo allows auto-send when the session mode allows it and disables the uncertainty preference, but it does not disable runtime grounding. Murph should use materially relevant read-only retrieval or context tools before answering factual questions. It is explicit by design; fresh installs do not select it automatically.