kingargyle

You are in PLANNING mode. You should take more time to research and think deeply about the given task, which will usually be more complex. Instead of bouncing ideas back and forth with the user very frequently, PLANNING mode means you should interact less often and do as much work independently as you can in between interactions. As much as possible, you should also verify your work before presenting it to the user.

You will be maintaining a number of artifact files that are useful in completing your task or documenting information relating to your task. These can be any file type, but the most common artifact type is an artifact document which should have a filename as .artifact.md. Your primary way of getting the user's feedback is to request review on these documents. All artifacts should be written to /home/kingargyle/.cache/Google/AndroidStudio2025.3.4/projects/q.a329d322/.artifacts/20260526-142738-70d48219-fa6c-4b78-89dd-59de15f7e4c9. You do NOT need to create this directory y

kingargyle

                       # SYSTEM PROMPT: CORE DIRECTIVE
                        You are an autonomous software engineer. Every architectural choice, test, and line of code
                        must strictly adhere to this Prime Directive. You have tools and you know how to use them.
                        
                        ## THE FOUR PILLARS
                        1. **Clean Architecture:** Strict separation of concerns. Keep business/domain logic
                           decoupled from UI, frameworks, and databases. Use interfaces at boundaries.
                        2. **Clean Code:** Prioritize readability, SOLID principles, and intent-revealing names.
                           Functions must be small, focused, and free of side effects.

3. Pragmatic Programming: No premature abstraction or over-engineering. Eliminate

kingargyle

AGENTS.md

Behavioral guidelines to reduce common LLM coding mistakes. Merge with project-specific instructions as needed.

Tradeoff: These guidelines bias toward caution over speed. For trivial tasks, use judgment.

1. Think Before Coding

Don't assume. Don't hide confusion. Surface tradeoffs.

kingargyle

name	skill-validator
description	Strict validator for auditing skills against the agentskills.io specification (directory structure, YAML, progressive disclosure, and best practices).

Skill Validator Standards

Purpose

This skill provides the logic required to audit and enforce compliance for ANY agent skill within the serenity-android project according to agentskills.io standards.

kingargyle

The following is part of the Agent's system level/developer level prompt that is sent from the Android Studio Panda 3 agent, to whatever LLM it is working with. This is new and not the same as prior information. This information is only sent if it can discover skills at either the .skills or .agent/skills locations. They have taken an approach where the skill discovery process, selection of when to use the skill, and the implementation, is not a TOOL like CoPilot or Claude Code, but a system level prompt.

This has advantages in particular ways.

1. It works across LLM models
2. It is really functionally agent agnostic besides the intial injection and discovery of skills by the Agent to provide to the LLM.
3. It allows for easier evolution of the skills implementation. As the standard evolves, you change the prompt.

kingargyle

Hidden Abilities of the Opencode Agent

Underneath the "magic" of an autonomous AI agent, the functionality is driven by a series of highly specialized, context-aware prompt-orchestration wrappers. My abilities—which often feel like autonomou s intelligence—are built upon several core "prompt-based" mechanisms:

1. The "TodoWrite" Manager

This is a structured prompt-management system. It forces me to maintain a specific, persistent list format to track "in-progress" states across multiple conversation turns. Without this, I would lose tr ack of my own progress.

2. The "Code-Reviewer" Subagent

kingargyle

[SPEC] Agent Skills Integration (agentskills.io Standard)

Goal

Implement the agentskills.io standard into the project's Phased Spec-Driven Development (SDD) protocol. This enables modular, pull-on-demand expertise discovery and activation, ensuring context economy while providing specialized agent capabilities using a universal root .skills directory.

Context

• Constraints: Must work with non-native skill loaders (Android Studio Gemini/Otter), Claude Code, OpenHands, and Opencode.
• Memory Path: prompts/plans/agent_skills_integration/memory/
• Files:
• prompts/plans/sdd_implementation_v1.md (Target for update)

kingargyle

Using the protocol in prompts/sdd_implementation_v1.md as a base, perform a "Phase 0: Agent Skills Integration" update to that file.

Update the document to Version 2.2.0 with the following mandatory architectural changes to support the agentskills.io standard:

1. FORMAL AGENT SKILLS DISCOVERY (v2.2.0):
   • Protocol Update: Define a specific "Skill Discovery" phase for agents that do NOT natively support external skill loaders (e.g., Android Studio Otter, Gemini, Copilot).
   • Storage Standard: All skills must reside in prompts/skills/<skill-name>/.
   • File Structure: Each folder MUST contain a single SKILL.md file (the interface) and optional scripts/ or references/ directories.
   • The Discovery Step: At the start of every session, the agent MUST execute list_files on prompts/skills/ to identify available expertise.

kingargyle

[SPEC] Agent Tool and Capability Discovery

Goal

• To systematically discover and document all features, tools, MCP servers, sub-agent support, skills (agentskills.io), and commands that the current agent environment reports to the LLM.
• Generate structured reports for each discovered category in the discovered_artifacts directory.

Context

• Current State: The agent has access to a variety of tools (standard and MCP-based), but there is no centralized documentation of these capabilities within the project for reference.
• Constraints: Must follow SDD v2.0 protocol. Reports must be generated without violating "Reporting Constraints" (except as requested for this specific discovery task).
• Files:
• prompts/plans/agent_tool_discovery/agent_tool_discover.md (This spec)

kingargyle

Write me a spec that can be used to test out the functionality of the  @sdd_implementation_v1.md  or any general SDD implementation. The spec should generate a test suite that can be run to validate the functions of the spec implementation is working as it is designed or intended. It should be able to update whatever existing SDD test suite exists if the underly spec implementation changes.

License

MIT License Copyright (c) 2026 David Carver and NineWorlds