Skip to content

Capability Portability Ranking

Last updated: 2026-03-15 Audit sources: PARTNER_OUTPUT_PORTABILITY_AUDIT.md, AI_AGENTS_AUDIT_AND_MIGRATION_PLAN.md

Strategic Context

This ranking evaluates all 40 agent capabilities against output portability and agent-readiness — the two dimensions that determine whether a third-party integrator (Tackle, Suger, Claazar, Labra) can programmatically consume Vellocity's output.

The Integration Pull Strategy

The goal is not to push integrators to adopt Vellocity. The goal is to make Vellocity outputs so clean, versioned, and machine-readable that partners pull integrators toward us:

"Hey Tackle — you need to integrate with Vellocity so I can improve my listing quality. You don't have these capabilities and it would take you a long time to build them."

This works when: 1. Output is portable — JSON with formal schema, not HTML blobs or PHP arrays 2. Output is discoverable — jsonSchema() method for MCP clients, OpenAPI compatibility 3. Output is versioned — schema_version field so integrators can pin to stable contracts 4. Output is self-describing — forOutputType() method returns context-appropriate views

Scoring Criteria

Grade Portability Agent-Readiness
A+ Formal schema class, versioning, toJson/fromJson/jsonSchema(), REST API, MCP discovery External agent can discover → inspect → execute → consume without Vellocity-specific knowledge
A Formal schema class, versioning, toJson/fromJson/jsonSchema() External agent can consume output with schema reference only
A- Structured JSON output, consistent keys, but no formal schema class Agent can parse reliably but must hardcode field expectations
B+ Structured JSON but no versioning, no validation, no schema discovery Agent can parse with documentation but no contract guarantees
B JSON output but inconsistent structure across operation types Agent needs per-type parsing logic
B- Mixed formats (arrays, strings, platform-specific) Agent needs significant transformation
C Human-readable output (Markdown, PDF) Agent cannot reliably consume

Tier 1: Platform Differentiators (A / A+)

Capabilities with formal schema versioning, MCP discovery, and high strategic uniqueness. These are the outputs integrators cannot replicate.

Rank Capability Slug Motion Port. Agent Credits Schema Class Strategic Value
1 Marketplace Listing Optimizer marketplace_listing_optimizer LEARN A A 25 MarketplaceScoreSchema v1.0 3 MCP tool schemas, export endpoint, 3 operation types. The gold standard.
2 AWS Blog Co-Author aws_blog_co_author CREATE A A 35 AwsBlogSchema v1.0 Encodes AWS bar-raiser rubric. Section ownership model (AWS/partner/co-authored). Unique.
3 Agent Tool Card agent_tool_card CREATE A A+ 25 AgentToolCardSchema v1.0 toMcpToolDefinition() + toOpenApiOperations(). Generates interop-ready specs.
4 Funding Application Writer funding_application_writer CREATE A A 30 — (structured JSON) Maps 1:1 to 5 AWS programs. Output directly submittable.
5 GTM Content (via Content Series) generate_content_series CREATE A A 30 GtmContentSchema v1.2 Universal GTM output schema. Crown jewel per portability audit.
6 First Call Enrichment — SUPPORT A A — FirstCallEnrichmentSchema v1.0 Multi-source research dossier. Persona-aware output views.
7 Joint GTM Planner joint_gtm_planner LAUNCH A A+ 40 — (workflow templates) Maps to Workflow Templates & Triggers (A/A+ in portability audit).
8 ACE Opportunity Sync ace_opportunity_sync CO-SELL A A 15 — (structured JSON) Output directly submittable to AWS ACE. Platform-native portability.

Tier 1 Integrator Value

These capabilities produce outputs that Tackle, Suger, Labra, and Claazar cannot generate and would take 6-12 months to build: - Listing Optimizer: No marketplace management tool scores listings with AI visibility + backlink authority + content quality in a single schema - Blog Co-Author: No tool encodes AWS editorial rubric with section ownership - Agent Tool Card: Generates MCP + OpenAPI specs — integrators consume these directly - Funding Writer: Maps to 5 specific AWS programs with field-level constraints

Tier 2: Schema Upgrade Candidates (B+ → A)

Capabilities with structured JSON output that need formal schema classes to reach A grade. These are the highest-ROI upgrades.

Rank Capability Slug Motion Current Target Credits Upgrade Path
9 RSS Post Scoring rss_post_scoring LEARN A- A 20 Add RssInsightSchema — formalize 4-dimension scoring + enrichment prompts
10 Marketplace Awareness marketplace_awareness LEARN B+ A 15 Add MarketplaceAwarenessSchema — formalize 6 analysis types with unified envelope
11 Competitor Analysis competitor_analysis LEARN B+ A 12 Add CompetitorAnalysisSchema — formalize 5 analysis types with versioning
12 Enrich Brand Voice enrich_brand_voice CREATE B A 8 Add BrandVoiceSchema — replace PHP array storage with versioned schema
13 Deal Influence Tracking deal_influence_tracking LEARN B+ A 15 Add DealInfluenceSchema — formalize 8 tracking types with JSON export
14 Content Readiness Analyzer content_readiness_analyzer LEARN A- A 10 Already structured; add jsonSchema() + forOutputType() methods
15 CPPO Proposal Generator cppo_proposal_generator CO-SELL A- A 25 Has SimulatableCapabilityInterface; add standalone schema versioning

Tier 2 Integrator Value

These capabilities are unique to Vellocity but currently require integrators to reverse-engineer the output format. Adding schemas means: - Tackle can consume listing quality signals to enhance their listing management - Suger can pull RSS intelligence scores to surface relevant content in their dashboards - Labra can ingest competitor analysis to enrich their marketplace positioning features - Claazar can consume brand voice data to personalize their co-sell workflows

Tier 3: Solid Capabilities (B+ steady state)

Structured output, valuable to users, but either (a) output is content-centric (not data-centric) or (b) schema formalization has lower ROI.

Rank Capability Slug Motion Port. Agent Credits Notes
16 Co-Sell Partner Matching cosell_partner_matching CO-SELL A- A- 20 Maps to Co-Sell Matching Simulation (A/A in audit). Good structure.
17 AWS Innovation Monitor aws_innovation_monitor LEARN A- B+ 15 Discovers AWS updates across 3 feeds. Structured but advisory.
18 Productization Recommendation productization_recommendation SUPPORT A- B+ 15 SA-level recommendations. Structured but advisory.
19 Marketplace Listing SEO marketplace_listing_seo LEARN A- B+ 25 AI assistant mention tracking. Novel but niche.
20 Partner Intelligence partner_intelligence CO-SELL A- B+ 10 Strength scores, warm intros. Good JSON but no schema.
21 SEO Intelligence seo_intelligence LEARN B+ B+ 10 Traffic, keywords, tech stack. External-data dependent.
22 Predict Partner Success predict_partner_success CO-SELL B+ B+ 15 Predictive/advisory output.
23 Content Gap Analysis content_gap_analysis LEARN B+ B+ 15 Keyword gap analysis. Externally sourced.
24 LinkedIn Graph Analysis linkedin_graph_analysis LEARN B+ B+ 10 Platform-specific constraints limit portability.
25 AWS Clean Rooms aws_cleanrooms_analysis CO-SELL B+ B+ 20 Depends on external AWS Clean Rooms integration.
26 Assess Launch Readiness assess_launch_readiness LEARN B+ B+ 10 Composite assessment. Shares handler with predict_performance.

Tier 4: Content Generation (B / B+)

Output is primarily content (text, images) rather than structured data. Schema formalization has minimal integrator value — integrators consume the content, not the metadata.

Rank Capability Slug Motion Port. Agent Credits Notes
27 Generate Text generate_text CREATE B+ B+ 10 Core generation. Content blob output.
28 Refine Content refine_content CREATE B+ B 10 Iterative. Needs context chain.
29 SEO Content Optimize seo_content_optimize CREATE B+ B+ 12 Rewrites for SEO.
30 SEO Content Analysis seo_content_analysis LEARN B+ B+ 8 17-factor grading.
31 Generate Meta Tags generate_meta_tags CREATE B+ B+ 5 Simple structured output.
32 Analyze Content analyze_content LEARN B+ B 5 Quality scores. Advisory.
33 Keyword Research keyword_research LEARN B B 5 Ephemeral data.
34 Discover Search Questions discover_search_questions LEARN B B 3 List output.
35 Predict Performance predict_performance LEARN B B+ 8 Advisory predictions.
36 Generate Image generate_image CREATE B B- 20 Binary output.
37 Publish to Social Media publish_to_social_media LAUNCH B+ B+ 0 Side-effect capability.

Tier 5: Infrastructure / Plumbing

Not ranked on portability — these enable other capabilities rather than producing standalone output.

Rank Capability Slug Motion Credits Notes
38 Query Knowledge Base query_knowledge_base SUPPORT 2 Retrieval tool. Essential plumbing.
39 Query Platform Knowledge query_platform_knowledge SUPPORT 2 Platform docs retrieval.
40 Fetch External URL fetch_external_url SUPPORT 2 Content fetcher utility.
— Sync Marketplace Listings sync_marketplace_listings SUPPORT 0 Import tool. Side-effect.

Schema Upgrade Implementation Plan

Priority 1: Highest Integrator Impact (This Sprint)

These 5 schemas unlock the most integrator value because they formalize capabilities that are unique to Vellocity and impossible for marketplace management tools to replicate quickly.

Schema Capability Key Benefit for Integrators
RssInsightSchema rss_post_scoring Tackle/Suger can surface Vellocity's relevance scores in their dashboards
MarketplaceAwarenessSchema marketplace_awareness Integrators can consume GTM signals without building their own news analysis
CompetitorAnalysisSchema competitor_analysis Integrators can enrich their positioning features with Vellocity's competitive data
BrandVoiceSchema enrich_brand_voice Integrators can read brand context to personalize their own outputs
DealInfluenceSchema deal_influence_tracking Integrators can correlate their deal data with Vellocity's attribution signals

Priority 2: Strengthen Existing (Next Sprint)

Schema Enhancement Capability Upgrade
Add jsonSchema() content_readiness_analyzer Formalize MCP discovery
Add standalone schema cppo_proposal_generator Version the SimulatableCapabilityInterface output
Add jsonSchema() cosell_partner_matching Formalize matching output

Implementation Pattern

Each schema follows the established MarketplaceScoreSchema pattern: 

class ExampleSchema {
    public const SCHEMA_VERSION = '1.0';

    // Typed properties
    public array $metadata;
    public ?array $analysis;

    // Required methods
    public function toArray(): array;
    public function toJson(): string;
    public static function fromJson(string $json): self;
    public function validate(): bool;
    public function getValidationErrors(): array;
    public function forOutputType(string $type): array;
    public static function jsonSchema(): array;
    public static function getEmptySchema(): array;
}

Integrator Consumption Model

How Tackle/Suger/Labra/Claazar Would Use These Schemas

Partner configures Vellocity agent
    ↓
Agent runs capability (e.g., marketplace_listing_optimizer)
    ↓
Output stored as versioned schema (MarketplaceScoreSchema v1.0)
    ↓
Integrator calls MCP tool or REST API
    ↓
Discovers schema via jsonSchema() endpoint
    ↓
Consumes structured output (toJson())
    ↓
Renders in their own UI or feeds into their workflows

What Makes This Defensible

  1. Schema versioning — Integrators can pin to v1.0 and not break when we ship v1.1
  2. MCP discovery — jsonSchema() method means integrators auto-discover output format
  3. Output views — forOutputType('agent') returns agent-optimized format vs. forOutputType('dashboard') for human display
  4. Validation — validate() method means integrators can verify output integrity before processing
  5. Domain expertise baked in — The schema encodes Vellocity's unique knowledge (AWS rubrics, scoring dimensions, ICP alignment weights) that integrators cannot easily replicate

The Partner's Pitch to Integrators

"I use Vellocity to score my marketplace listing (82/100, B+ grade), analyze competitor positioning, and track which content drives pipeline. Your tool manages my listing but can't tell me why it's underperforming. Vellocity outputs everything as versioned JSON with formal schemas — just call their API and render the scores in your dashboard. It's a 2-week integration, not a 2-year build."