configuration-management¶
Purpose¶
Provides structured, data-driven storage for AI agent configurations that remain tool-agnostic while enabling generation of tool-specific outputs. Maintains single source of truth for commands, agents, and related tooling across multiple AI development environments.
Requirements¶
Requirement: Data-Driven Repository Structure¶
The repository SHALL organize AI tool configurations as structured data sources with 3-tier separation to enable tool-agnostic maintenance while generating tool-specific outputs.
Scenario: Skills as a first-class item type¶
WHEN reusable expertise is added to the repository
THEN it can be represented as “skills” with:
tool-agnostic metadata
a tool-agnostic body by default (with optional tool-specific overrides)
templates that emit tool-compatible frontmatter and discovery paths
Scenario: Tool-specific permissions mapping¶
WHEN a skill specifies “allowed tools” in tool-agnostic semantic form
THEN the system maps/format those tool specifications per coder where supported
AND the system omits or warns for fields that are unsupported by a given target tool (without failing generation)
Requirement: Structured Source Data Management¶
The system SHALL maintain structured source data that generates content for multiple AI tools from single sources while supporting diverse tool formats.
Priority: Critical
Scenario: Single source of truth¶
WHEN command or agent is defined
THEN metadata stored in tool-agnostic TOML configurations
AND single source drives generation for all supported AI tools
Scenario: Content body separation¶
WHEN tool-specific content is needed
THEN content bodies separated by coder
AND appropriate fallback strategies applied
Scenario: Format handling¶
WHEN different AI tools require different formats
THEN generic templates handle format differences
AND transformations applied at generation time
Scenario: Hook script distribution¶
WHEN hook scripts are distributed
THEN distributed via minimal Copier template
AND not dynamically generated