Agent Ex Phase 0: Add agent-asset convention and authoring docs

[brooke-hamilton]

Area for Improvement

Add foundational contributor documents: contributing-agent-assets.md, authoring-contributing-docs.md, and extending-agent-ex.md to the docs/contributing/ folder. These will encode file-strategy, naming, and contribution conventions for the Agent Ex system, following Phase 0 — Meta-tooling foundation of the plan.

• contributing-agent-assets.md: File-strategy rule, file-size budgets, CI gates, naming conventions, templates for AGENTS.md/instructions/skills/prompts/agents.
• authoring-contributing-docs.md: Standard formats for contributor and architecture docs (Purpose→Prereqs→Steps→Verification→Troubleshooting; Entry points→Packages→Flow→Change-safety).
• extending-agent-ex.md: "Add a capability" decision tree, repo-onboarding checklist.
• (Optionally) Add a stub README.md in the same folder.

Docs should be stand-alone and reference the plan. No code/tooling/skills are added in this PR; it's pure doc work.

Observed behavior

No foundation for consistent contributor/agent docs and conventions. No central place for the file-strategy, naming rules, etc. Later phases are blocked without these docs.

There are existing instructions and skills, but no doc specifying how to add new ones or the conventions to follow.

Desired behavior

Deterministic (CI):

• Markdown lint, link check, and cspell pass over the new docs.
• A path-existence check confirms every repo path referenced in the docs resolves (no hallucinated/nonexistent paths).
• The stub README.md links to all three docs.
• Embedded code or template blocks inside the docs parse as valid YAML, Markdown, or Shell script as relevant.

Content:

• authoring-contributing-docs.md follows the format it prescribes.
• extending-agent-ex.md covers all five asset types (doc/instruction/skill/prompt/agent) plus the onboarding checklist.
• Naming conventions match existing .github/skills and .github/agents.
• CI-gate list matches what is (or will be) enforced with repo automation.

Behavioral:

• An agent given only extending-agent-ex.md can scaffold a correctly-named, correctly-placed new skill with a backing-doc link, with no extra guidance.
• An agent given authoring-contributing-docs.md and a topic produces a draft in the expected format.
• An agent given only extending-agent-ex.md can execute the onboarding checklist for a new repo.

Proposed Fix

Draft initial versions of all 3 docs leveraging the conventions, templates, and process outlines in Phase 0 — Meta-tooling foundation for structure, and reference this spec in the docs. No implementation/code included. Link from stub README.md. Optionally, add a PR test script for path and link existence checks that will be reused as a CI gate.

rad Version

edge

Operating system

No response

Additional context

No response