Add a dedicated Guideline 4.3(b) human-craft evaluation

[ElxMaj]

What changed

• Replaces the generic visual-report scoring scenario with a focused Mode C human-craft audit.
• Adds realistic evidence for a template-derived freelancer focus timer.
• Scores the exact five-grade contract, current 4.3(b) interpretation, template provenance, product depth, unverified surfaces, and five actionable interventions.
• Makes the rejection-recovery task explicitly invoke the installed skill for a more stable with-skill comparison.
• Adds the live Maj Labs Tessl quality and impact badge to the README while keeping the three-badge limit.
• Prepares plugin manifests for version 1.1.6.

Why

The previous renderer scenario tested a useful feature but did not clearly separate specialized App Store review judgment from general HTML generation. The new scenario measures the skill's strongest differentiator: evidence-based analysis of the symptoms Apple may reject under the June 2026 Guideline 4.3(b) wording, without claiming AI detection or recommending cosmetic disguise.

The visual report renderer remains part of the skill and its tests remain in place. Only the public Tessl evaluation scenario changes.

Validation

• 12 Python tests passed
• Skill package validation passed
• Tessl plugin lint passed
• Tessl publish dry run with scenario-quality checks passed
• New weighted rubric totals 100 points
• Tessl badge endpoint returned the current 95% (↑1.67x) result

Release gate

Keep this PR in draft until enough Tessl credits are available to run all three scenarios. Merging to main triggers the Tessl publish workflow, so public version 1.1.4 should remain the stable release until version 1.1.6 has a complete comparison.