Manage safety rule packs and block/allow policies for AI harnesses.
npx skills add https://github.com/AesopScott/mojo --skill safety-packs
Please help me install this Agent Skill. Skill: safety-packs Source: https://github.com/AesopScott/mojo/tree/main/harnesses/skills/safety-packs Command: npx skills add https://github.com/AesopScott/mojo --skill safety-packs
This skill provides a structured framework for managing safety enforcement, allowing operators to define, test, and implement granular capability controls and block/allow policies without accidental overreach.
Use the safety-packs skill to inspect the current safety harness and propose a new block policy for the specified capability boundary.
High-intent search queries and answers about installing and using this skill.
AI safety boundaries are enforced by managing safety rule packs and block/allow policies to govern capability controls, preventing models from executing unauthorized actions or accessing restricted APIs.
A safety rule pack is a structured framework defining specific block or allow policies for AI actions. You need it when deploying new AI features to ensure strict adherence to operational risk mitigation boundaries.
You configure block and allow policies by defining granular capability controls within the safety harness, explicitly blocking sensitive internal APIs while permitting interaction with approved public-facing resources.
Yes, safety rule packs apply directly to system configuration auditing and risk mitigation workflows, requiring precise definition of failure modes, control levers, and verification tests to ensure safe operational changes.
The best way to define failure modes and control levers for AI governance is using a structured safety pack framework to propose, inspect, and test block policies against specified capability boundaries.
Without precise verification tests, managing AI safety policies risks accidental overreach or boundary failures, as operators cannot ensure safe operational changes or validate that capability controls function correctly.
AesopScottConfigure risk-tiered proof requirements and waiver shortcuts in the Proof Authority harness.
davidahmannTest Gait policy changes against intent fixtures and simulate staged rollouts.
AesopScottBind objective types to goal harnesses using schema constraints.
jaskaranhundalAuthor OPA/Rego security policies and assess control effectiveness against NIST, ISO 27001, and CIS Controls.
