Standardize experimental findings format with a contribution template

[olaservo]

Background

Experimental Findings contains valuable implementation results but the entries vary in structure. McpGraph has findings/resolution/concerns, worfcat's comparison uses a numbered list with results per approach, and others are brief paragraphs. As more contributors add findings (#37, #38, #40, #41, #42), a consistent template will make entries comparable and ensure nothing important is omitted.

Scope

1. Design a findings template covering:

   • Implementation: Name, repo URL, author
   • Approach tested: Which approach(es) from docs/approaches.md does this map to?
   • Setup: Clients tested, models tested, configuration notes
   • What was tested: Specific scenarios or behaviors evaluated
   • Results: What worked, what didn't, what was surprising
   • Requirements addressed: Which requirements (from docs/requirements.md) does this provide evidence for?
   • Limitations: What remains untested or uncertain
   • Date: When testing was performed (relevant as clients and models evolve)

2. Add the template to either:

   • A new section at the top of docs/experimental-findings.md, or
   • A separate docs/findings-template.md referenced from CONTRIBUTING.md

3. Retrofit at least 2 existing entries to demonstrate the format (McpGraph and worfcat's comparison are good candidates).

4. Update CONTRIBUTING.md with guidance on using the template when adding findings.

Acceptance Criteria

• Template documented and accessible
• At least 2 existing findings retrofitted to the template
• CONTRIBUTING.md updated with instructions for submitting findings
• Template reviewed by at least 1 facilitator

References

• Experimental Findings
• CONTRIBUTING.md
• Research issues that will produce new findings: Compare skill delivery mechanisms: file-based vs MCP-based #37, Survey client resource-loading support across major MCP clients #38, Build client-side reference implementation for model-driven resource loading #40, Build server-side reference implementation for skills-as-resources #41, Test skills-via-sampling approach (Approach 3 variant) #42