fix(commands): rename NFR references to success criteria in analyze and clarify

[ismaelJimenez]

Description

Replaces "Non-Functional Requirements" terminology with "Success Criteria" in the analyze and clarify command templates to fix a mismatch with the spec template.

The problem:

The analyze command instructs the agent to load a "Non-Functional Requirements" section from specs — but that section is not part of the spec template. The specify command generates specs using spec-template.md, which defines a "Success Criteria" section (with "Measurable Outcomes" underneath), not a "Non-Functional Requirements" section. This means analyze was looking for a section name that doesn't match what the spec actually contains, forcing the agent to infer mappings or miss the Success Criteria content when assessing coverage gaps.

The only way NFR-like content could appear in a spec was through the clarify command, which uses its own label ("Non-Functional Quality Attributes") that doesn't match the spec template's actual section name either. This meant non-functional content could end up scattered — partially under "Success Criteria > Measurable Outcomes" and partially under a clarify-only "Non-Functional / Quality Attributes" heading — and only when a user's clarification response happened to surface a non-functional constraint. In the typical workflow (specify → analyze), these labels never align.

What this PR does:

• analyze.md: Points spec-loading at "Success Criteria" instead of the nonexistent "Non-Functional Requirements" section. Refines the requirements inventory to include only Success Criteria items that require buildable work (e.g., load-testing infrastructure, security audit tooling), explicitly excluding post-launch outcome metrics and business KPIs that aren't actionable as tasks.
• clarify.md: Updates the clarification routing rule to direct non-functional constraints to "Success Criteria > Measurable Outcomes" — matching the actual section name in the spec template.

Testing

• Tested locally with uv run specify --help
• Ran existing tests with uv sync --extra test && uv run pytest (435 passed)
• Tested with a sample project

AI Disclosure

• I did not use AI assistance for this contribution