feat: Persona-to-Artifact Mapping

[katriendg]

Parent: #431

Issue Description

Define the final persona-to-artifact mapping that determines which agents, prompts, instructions, and skills belong to each persona collection. This issue captures the team discussion and consensus needed before artifacts can be tagged in the registry.

Motivation

The registry schema (Phase 1) provides the structure, but the actual assignments require human judgment. Different team members have perspectives on which artifacts serve which roles. Getting this mapping right determines the usefulness of every persona-targeted package. This issue is the forum for that discussion.

Decisions Required

1. Persona Taxonomy Validation

Confirm or adjust the proposed persona definitions:

Persona ID	Display Name	Target Audience
developer	Developer	Software engineers writing code
tpm	Technical PM	Program/product managers
devops	DevOps Engineer	Platform, SRE, infrastructure engineers
architect	Software Architect	Solution/system architects
technical-writer	Technical Writer	Documentation specialists

Note: A data-scientist persona is under evaluation. The HVE Data Science extension already covers this audience. This persona may be added later depending on how the two extensions relate.

Open questions:

• Are these the right persona boundaries?
• Should any personas be merged or split?
• Are the display names and descriptions accurate?

2. Universal Artifacts

Identify which artifacts belong in every collection regardless of persona. Candidates include the core RPI workflow agents and shared instructions, but the final list needs team agreement.

3. Per-Persona Artifact Assignments

For each persona, determine which non-universal artifacts should be included. Each artifact (agent, prompt, instruction, skill) needs at least one persona tag. Artifacts can belong to multiple personas.

4. Agent Dependency Validation

Confirm the agent-to-prompt and agent-to-instruction dependency mappings. When an agent is installed, its required prompts and instructions must be included automatically regardless of persona filter.

Process

1. Post initial proposed mapping as a comment on this issue
2. Team reviews and provides feedback via discussion
3. Iterate on assignments until consensus is reached
4. Apply final mapping to ai-artifacts-registry.json

Acceptance Criteria

• Persona taxonomy confirmed or adjusted by team
• Universal artifact set agreed upon
• Every artifact has at least one persona assignment
• Agent dependency mappings validated (no agent without its required prompts/instructions)
• Final mapping applied to ai-artifacts-registry.json
• Validation passes with no untagged artifacts

Additional Context

• Persona IDs use kebab-case: developer, tpm, devops, architect, technical-writer
• Current artifact counts: 21 agents, 22 prompts, 17 instructions, 1 skill
• Depends on: feat: AI Artifacts Registry Design and Schema #432 (Phase 1: Registry Design and Schema) for the registry file structure
• Feeds into: feat: Multi-Package Extension Build System #433 (Phase 2: collection manifests), feat: Installer Agent Persona Selection and All-Artifact Copying #434 (Phase 3: installer filtering), feat: Documentation Updates for Persona Distribution #435 (Phase 4: documentation)

[katriendg]

This issue has been superseded by the collection-based distribution system implemented in #519 (PR #439). The persona taxonomy and central registry mapping are no longer needed — collections directly list their items by path and kind in YAML manifests. All 10 domain collections are defined with complete item lists.

Closing as completed in favor of #519.