KS
Killer-Skills

homelab-durability — how to use homelab-durability how to use homelab-durability, homelab-durability setup guide, homelab-durability alternative, homelab-durability vs other AI agent skills, what is homelab-durability, homelab-durability install, homelab-durability configuration, homelab-durability sources of truth, homelab-durability execution contracts

v1.0.0
GitHub

About this Skill

Perfect for AI Agents needing reliable homelab operation and durability, particularly those utilizing operating models and execution contracts. homelab-durability is a technical framework that ensures reliable operation of AI agents through strict execution contracts and conservative defaults.

Features

References sources of truth from docs/_core/OPERATING_MODEL.md and docs/_core/SOURCES_OF_TRUTH.md
Enforces strict execution contracts for reliable operation
Implements conservative defaults, including a 500MB download approval threshold
Requires disk checks before model downloads or large actions
Supports topology configuration via docs/foundation/topology.md
Provides a PLATFORM_DOSSIER for detailed documentation

# Core Topics

bebopbabailey bebopbabailey
[0]
[0]
Updated: 3/6/2026

Quality Score

Top 5%
44
Excellent
Based on code quality & docs
Installation
SYS Universal Install (Auto-Detect)
Cursor IDE Windsurf IDE VS Code IDE
> npx killer-skills add bebopbabailey/homelab-llm/homelab-durability

Agent Capability Analysis

The homelab-durability MCP Server by bebopbabailey is an open-source Categories.community integration for Claude and other AI agents, enabling seamless task automation and capability expansion. Optimized for how to use homelab-durability, homelab-durability setup guide, homelab-durability alternative.

Ideal Agent Persona

Perfect for AI Agents needing reliable homelab operation and durability, particularly those utilizing operating models and execution contracts.

Core Value

Empowers agents to ensure reliable operation by providing sources of truth, such as OPERATING_MODEL.md and SOURCES_OF_TRUTH.md, and enforcing strict execution contracts with conservative defaults, including disk checks and download approval thresholds, utilizing Markdown documentation and layer-gateway constraints.

Capabilities Granted for homelab-durability MCP Server

Enforcing strict execution contracts for reliable homelab operation
Managing sources of truth for consistent decision-making
Configuring conservative defaults for safe model downloads and caching

! Prerequisites & Limits

  • Requires adherence to specified operating models and change rules
  • Needs Markdown documentation for sources of truth
  • Imposes a download approval threshold of 500MB and requires disk checks for large actions
Project
SKILL.md
3.0 KB
.cursorrules
1.2 KB
package.json
240 B
Ready
UTF-8

# Tags

[No tags]
SKILL.md
Readonly

Homelab Durability

Sources of Truth (Must read first)

  • docs/_core/OPERATING_MODEL.md
  • docs/_core/SOURCES_OF_TRUTH.md
  • docs/_core/CHANGE_RULES.md
  • docs/foundation/operating-rhythm.md
  • layer-gateway/CONSTRAINTS.md
  • docs/foundation/topology.md
  • docs/PLATFORM_DOSSIER.md

Execution Contract (Strict)

Defaults (Conservative)

  • Download approval threshold: 500MB.
  • Disk check required before any model download/cache action or any explicitly large (>5GB) action: YES.
  • Host confirmation: strict for ops commands, but may be treated as sticky within the same session after explicit confirmation.
  • Violations: HARD-BLOCK for MLX/disk/host/file-list rules; WARN for stylistic rules.

Workflow Rules (Behavioral)

  • Use SCRATCH_PAD.md only when the user asks. This is mainly for use if user needs to run commands themselves or asks to review commands/outputs/diffs; otherwise inline is acceptable.
  • NOW.md must reflect the active development task; update it when the active work changes.
  • NOW.md contains only active work + a single “NEXT UP” section. Everything else goes to BACKLOG.md.
  • Journal entries are append-only. Do not move or delete files in docs/journal/. If a correction is needed, add a new dated entry and update docs/journal/index.md.

Required startup declarations

Every response MUST begin by stating:

  1. Host (Mini vs Studio vs Orin)
  2. Scope (layer/service)
  3. Files (explicit list of intended files to change)

Hard-block rules (must refuse)

  • No multi-file edits without an explicit file list.
  • No ops commands unless Host Check is explicitly satisfied in the same response.
  • No restarts/service modifications unless Stage = Validate and a rollback is specified.
  • No disk repair / fsck / apfs repair / partition changes without explicit user approval.
  • MLX backends: MUST use mlxctl + registry as source of truth. No ad-hoc edits or direct mlx-openai-server control unless explicitly authorized.
  • Any download/cache/model pull > 500MB requires explicit user approval.

Disk guardrail

  • Before any model download/cache action or any explicitly large (>5GB) action, require a disk free check command suggestion and STOP until user confirms.

Default stance

  • Default to read-only inspection first.

Host Check (required before ops commands)

  • Explicitly confirm target host and its role.
  • If not confirmed, STOP and ask for confirmation.

Preferred Operating Rhythm

Inventory → Constraints → Minimal contract → Wire → Validate → Codify → Prune

Output Format (Required)

Mode: Discover|Design|Build|Verify Stage: Inventory|Constraints|Minimal Contract|Wire|Validate|Codify|Prune Host: <explicit> Scope: <layer/service> Proposed Change: <one paragraph> Files: <explicit list> Commands: <copy/paste safe; no wrapped lines> Validation: <commands> Rollback: <commands>

How to use

  • Invoke by name: "use homelab-durability"
  • Behavior changes:
    • Host check gate before any ops commands
    • File list gate before edits
    • Validation gate requires rollback for restarts

Related Skills

Looking for an alternative to homelab-durability or building a Categories.community AI Agent? Explore these related open-source MCP Servers.

View All

widget-generator

Logo of f
f

widget-generator is an open-source AI agent skill for creating widget plugins that are injected into prompt feeds on prompts.chat. It supports two rendering modes: standard prompt widgets using default PromptCard styling and custom render widgets built as full React components.

149.6k
0
Design

chat-sdk

Logo of lobehub
lobehub

chat-sdk is a unified TypeScript SDK for building chat bots across multiple platforms, providing a single interface for deploying bot logic.

73.0k
0
Communication

zustand

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication

data-fetching

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication