Feature heauristic checks + cleanup

[nitink23]

What's the issues or discussion related to this PR ?

Finding a better way to add heuristic checks. Cleaned up all the unnecessary files and unused functionality. Fixed the bug with versioning specialist definitions so everytime a specialist is enriched the version also increases. You can also manually bump up the version of your specialist.

What's added in this PR?

scenario.yamls now contain a section for heursitic checks which can be commands run for code or pattern matching looking for sections of a PRD.

heuristic_checks:

  enabled: true

  # Critical validation commands

  commands:

    - name: "install_succeeds"

      command: "pnpm install"

      weight: 2.0

      description: "Dependencies install successfully"

    - name: "build_succeeds"

      command: "pnpm build"

      weight: 3.0

      description: "Production build completes without errors"

    - name: "lint_passes"

      command: "pnpm lint"

      weight: 1.0

      description: "Code meets linting standards"

  # Check for Client Component patterns

  patterns:

    - name: "has_use_client_directive"

      file: "app/**/*.tsx"

      pattern: "^['\"]use client['\"]"

      weight: 3.0

      description: "Client component has 'use client' directive"

    - name: "uses_state_hook"

      file: "app/**/*.tsx"

      pattern: "useState|const \\[.*,.*set.*\\]\\s*=\\s*useState"

      weight: 2.0

      description: "Uses useState for state management"

    - name: "has_event_handlers"

      file: "app/**/*.tsx"

      pattern: "onClick|onChange|onSubmit|on[A-Z]\\w+"

      weight: 1.5

      description: "Event handlers implemented for interactivity"

    - name: "has_effect_hook"

      file: "app/**/*.tsx"

      pattern: "useEffect"

      weight: 1.0

      description: "Uses useEffect for side effects (optional)"

    - name: "imports_react_hooks"

      file: "app/**/*.tsx"

      pattern: "import.*\\{.*useState.*\\}.*from.*['\"]react['\"]"

      weight: 1.5

      description: "Imports React hooks properly"

    - name: "has_typescript_types"

      file: "app/**/*.tsx"

      pattern: "interface|type\\s+\\w+.*="

      weight: 0.5

      description: "TypeScript types defined"

Along with the new enhancements, the template version has increased.

What are the steps to test this PR?

pnpm bench

Run 002-client-component in the next.js suite to view heuristic changes.

to test the version bumping enable enrichment as well.

Documentation update for this PR (if applicable)?

Documentation already exists in docs/heuristic-checks-guide.md which covers the heuristic checks feature comprehensively. The guide includes:

• Configuration examples
• Check types (commands, files, patterns, structured, scripts)
• Weighting system
• Scoring integration
• Practical examples

No additional documentation updates required as the existing guide already covers the new heuristic_checks section in scenario.yaml files.

(Optional) What's left to be done for this PR?

[] MCP + oauth should be created @Nsttt
[] make a easier way to create pattern matching heuristics
[] prompt and conversation should be sent R2 instead of D1 since the JSON's get really big for longer benchmarks

(Optional) What's the potential risk and how to mitigate it?

Who do you wish to review this PR other than required reviewers?

@Nsttt @zackarychapple

(Required) Pre-PR/Merge checklist

• I have added/updated our documentation to cover this new behavior
• I have added an explanation of my changes
• I have written new tests (if applicable)
• I have tested this locally (standing from a first time user point of view, never touch this app before)
• I have mentioned the related person or team responsible for reviewing proposed changes
• I have/will run tests, or ask for help to add test