bastion

Orchestrate multi-component features with autonomous agents under milestone gates.

Updated Feb 12, 2026
One-click install
npx skills add https://github.com/DalvinCodes/ai-skills --skill bastion
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: bastion
Source: https://github.com/DalvinCodes/ai-skills/tree/main/bastion
Command: npx skills add https://github.com/DalvinCodes/ai-skills --skill bastion

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables you to delegate entire features to autonomous Feature Agents, reducing manual orchestration and overhead by enforcing milestone gates, persistent memory, and hierarchical validation.

Core Features & Use Cases

  • End-to-end feature orchestration across components with gate-based validation.
  • Persistent memory and resumption via the .bastion/<session-id>/ directory.
  • Clear role separation: PE gates and Feature Agents with sub-agents for tasks.

Quick Start

Initiate Bastion to start a new feature orchestration; the system will create a session, spawn a Feature Agent, and begin Gate 1 exploration.

Frequently Asked Questions about bastion

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I orchestrate complex features across multiple components autonomously?

Autonomous feature orchestration is handled by delegating tasks to Feature Agents that coordinate cross-component implementation under milestone gates with persistent memory. This reduces manual overhead by automating exploration, planning, and hierarchical validation throughout the feature lifecycle.

What is gate-based validation for multi-component feature implementation?

Gate-based validation enforces milestone checkpoints during feature implementation, requiring centralized PE review before autonomous agents proceed. This hierarchical validation mechanism ensures cross-component coordination meets quality standards at each stage of exploration, planning, and handoff.

How do I resume an interrupted autonomous agent feature orchestration session?

Resuming interrupted feature orchestration sessions relies on persistent memory stored in the .bastion/<session-id>/ directory. This mechanism allows autonomous Feature Agents to recover previous exploration and planning state, enabling seamless resumption of multi-component implementation tasks.

When do I need autonomous agents for cross-component feature coordination?

Autonomous agents are needed for cross-component feature coordination when tasks require end-to-end orchestration involving exploration, planning, and implementation across multiple components. Complex features benefiting from milestone gates and hierarchical PE review are ideal candidates for this approach.

Does autonomous feature orchestration support role separation between planning and implementation?

Role separation is supported through distinct PE gates managing validation and Feature Agents executing implementation with sub-agents for specific tasks. This structure ensures autonomous feature orchestration maintains centralized review authority while delegating cross-component execution to specialized agents.

What are the limitations of using autonomous agents for end-to-end feature orchestration?

Limitations of autonomous feature orchestration include dependency on milestone gate validation, which may slow delivery if centralized PE review is delayed. Additionally, complex cross-component features requiring real-time human intervention may not fully benefit from persistent memory and autonomous handoff mechanisms.