feat: AI-native framework integration (v1.3.0)#1
Conversation
New data/ai-native-framework.md with 4 AI-native criteria, 5 bolted-on indicators, the removal test, Karpathy hierarchy, architecture patterns, and AI-native pricing models. CLAUDE.md updated to document the new file.
product-review gets a 6th scoring dimension (AI Architecture) with the removal test and 4-criteria evaluation. idea-validation gets AI Architecture Fit as a validation signal with higher-ed cross-referencing.
go-to-market adds usage-based pricing playbook. pitch-review coaches the "improves with models" narrative. sales-strategy adds AI-native objection handling. fundraising-guide targets different VC pools.
edtech-landscape maps AI-native competitors. evidence-check adds behavior change dimension. accessibility-check flags AI bias and transparency. pilot-design adds AI-specific metrics.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request integrates a comprehensive AI-native framework into the EdTech Founder Stack, upgrading the system to version 1.3.0. The changes include a new reference data file defining AI-native vs. bolted-on criteria and the systematic update of all ten skills to include AI posture detection, specialized metrics, and tailored strategic guidance for AI products. Review feedback correctly identifies several minor inconsistencies where specific mapping instructions or scoring rubrics were omitted in certain skill files compared to others, ensuring the framework operates uniformly across the repository.
| - AI is a minor or planned feature — optional, supplementary, or not yet built | ||
| - No AI component | ||
|
|
||
| If AI-native or borderline: read `data/ai-native-framework.md` and add "behavior change" as an evidence dimension in Phase 2. AI-native products need to demonstrate they create behavior change (users work fundamentally differently), not just engagement or satisfaction. |
There was a problem hiding this comment.
The mapping instructions for the AI posture detection question are missing in this skill. All other skills that implement this detection block include explicit mapping for the LLM to classify the product as 'AI-native' or 'Borderline'. Adding this ensures consistent behavior across the framework.
| If AI-native or borderline: read `data/ai-native-framework.md` and add "behavior change" as an evidence dimension in Phase 2. AI-native products need to demonstrate they create behavior change (users work fundamentally differently), not just engagement or satisfaction. | |
| Map: "AI IS the product" = AI-native. "Significant feature" = Borderline. "Minor/planned" or "No AI" = Skip AI-specific evidence guidance. | |
| If AI-native or borderline: read `data/ai-native-framework.md` and add "behavior change" as an evidence dimension in Phase 2. AI-native products need to demonstrate they create behavior change (users work fundamentally differently), not just engagement or satisfaction. |
| Read `data/ai-native-framework.md`. Evaluate: | ||
|
|
||
| - **Removal test:** Remove the AI from this idea. Does it still solve the problem? If yes, the AI is supplementary, not essential. | ||
| - **Model improvement trajectory:** Will this idea get better as base models improve? Or is the AI a static feature that won't evolve? | ||
| - **Behavior change potential:** Would users of this product work fundamentally differently, or just slightly faster? | ||
| - **Higher ed cross-reference:** If the founder targets higher ed, check `data/higher-ed-jobs-atlas.md`. Some jobs are naturally AI-native (adaptive tutoring, intelligent scheduling, personalized content generation). Others don't need AI (financial aid navigation, transfer credit mapping). Flag if the founder is forcing AI into a job that doesn't need it. |
There was a problem hiding this comment.
This dimension lacks the scoring rubric defined in the /product-review skill. Since the scorecard on line 186 expects a score for 'AI Architecture Fit', providing a consistent rubric (using even numbers 2-10) ensures the LLM evaluates the idea accurately and maintains consistency across the Founder Stack.
| Read `data/ai-native-framework.md`. Evaluate: | |
| - **Removal test:** Remove the AI from this idea. Does it still solve the problem? If yes, the AI is supplementary, not essential. | |
| - **Model improvement trajectory:** Will this idea get better as base models improve? Or is the AI a static feature that won't evolve? | |
| - **Behavior change potential:** Would users of this product work fundamentally differently, or just slightly faster? | |
| - **Higher ed cross-reference:** If the founder targets higher ed, check `data/higher-ed-jobs-atlas.md`. Some jobs are naturally AI-native (adaptive tutoring, intelligent scheduling, personalized content generation). Others don't need AI (financial aid navigation, transfer credit mapping). Flag if the founder is forcing AI into a job that doesn't need it. | |
| Read `data/ai-native-framework.md`. Evaluate the idea against the 4 AI-native criteria (Token scaling, Model improvement, Removal test, Behavior change). | |
| Score on the AI-native criteria (use even numbers to distinguish from the 1-10 scale of other dimensions): | |
| - **10:** All 4 criteria met. Architecture is model-agnostic. Clear token economics. Memory/personalization across sessions. | |
| - **8:** 3 of 4 criteria met. Strong AI integration but missing memory, behavior change, or model-agnostic design. | |
| - **6:** 2 of 4 criteria met. AI is meaningful but not load-bearing. The product is in a borderline zone. | |
| - **4:** 1 criterion met. AI is a significant feature but the product works without it. Bolted-on with potential. | |
| - **2:** 0 criteria met but product claims to be "AI-powered." Sparkle icons and a chat pane. | |
| **Higher ed cross-reference:** If the founder targets higher ed, check `data/higher-ed-jobs-atlas.md`. Some jobs are naturally AI-native (adaptive tutoring, intelligent scheduling, personalized content generation). Others don't need AI (financial aid navigation, transfer credit mapping). Flag if the founder is forcing AI into a job that doesn't need it. |
| - AI is a minor or planned feature — optional, supplementary, or not yet built | ||
| - No AI component | ||
|
|
||
| If AI-native or borderline: read `data/ai-native-framework.md` and add AI-specific pilot metrics to Phase 2. |
There was a problem hiding this comment.
The mapping instructions for the AI posture detection question are missing here. Including the explicit mapping for 'AI-native' and 'Borderline' ensures the LLM correctly triggers the AI-specific pilot metrics in Phase 2, maintaining consistency with the other skills in the repository.
| If AI-native or borderline: read `data/ai-native-framework.md` and add AI-specific pilot metrics to Phase 2. | |
| Map: "AI IS the product" = AI-native. "Significant feature" = Borderline. "Minor/planned" or "No AI" = Skip AI-specific pilot metrics. | |
| If AI-native or borderline: read `data/ai-native-framework.md` and add AI-specific pilot metrics to Phase 2. |
| **If Evidence Readiness scored < 7:** | ||
| "You need stronger evidence before institutional buyers will take you seriously. Run `/evidence-check` to assess your evidence tier and build a plan." | ||
|
|
||
| **If AI Architecture scored < 5 (and AI is involved):** |
There was a problem hiding this comment.
There is a minor inconsistency between the warning threshold defined in Phase 2 (line 162, score < 6) and the recommendation trigger here in Phase 5 (score < 5). While logically equivalent given the even-number scoring system (2, 4, 6, 8, 10), using < 6 here would be more consistent and clearer for maintenance.
| **If AI Architecture scored < 5 (and AI is involved):** | |
| **If AI Architecture scored < 6 (and AI is involved):** |
1 participant
Summary
Every skill is now AI-native aware. When a founder's product involves AI, skills detect whether the AI is load-bearing (AI-native) or decorative (bolted-on) and adapt guidance accordingly.
Foundation:
data/ai-native-framework.mdreference data: 4 AI-native criteria, 5 bolted-on indicators, removal test, Karpathy hierarchy, architecture patterns, pricing modelsHeavy integration (new scoring dimensions):
/product-review— 6th scoring dimension: AI Architecture (criteria-based evaluation)/idea-validation— AI Architecture Fit signal with higher-ed cross-referencingMedium integration (adapted guidance):
/go-to-market— Usage-based pricing playbook, model improvement positioning/pitch-review— "Improves with models" investor narrative, AI-native VC targeting/sales-strategy— AI-native objection handling, demo flow, procurement narrative/fundraising-guide— AI-native vs traditional VC targetingLight integration (specific additions):
/edtech-landscape— AI-native vs bolted-on competitive mapping/evidence-check— Behavior change evidence dimension/accessibility-check— AI bias, transparency, explainability concerns/pilot-design— AI-specific pilot metrics (accuracy, hallucination, trust)The framework is diagnostic, not prescriptive. Bolted-on AI can be a valid strategy.
Pre-Landing Review
No issues found. All changes are markdown files (SKILL.md, reference data, documentation).
Adversarial Review
8 issues auto-fixed (wording consistency across skills, trigger logic, unsourced claim, missing next-skill recommendation). 3 pre-existing issues noted but not blocking.
Plan Completion
16/16 plan items DONE. 0 PARTIAL, 0 NOT DONE.
Scope Drift
Scope Check: CLEAN
TODOS
Test plan
/product-reviewwith AI-native product description — verify AI Architecture dimension appears/idea-validationwith AI idea — verify AI Architecture Fit signal/go-to-marketwith AI product — verify usage-based pricing guidance🤖 Generated with Claude Code