Rudra Wiki

Cross-Project Connections

3 min read

Cross-Project Connections

Last Updated: 2026-04-14 (S55) | Source: /synthesizer + session observations

The links between projects that are invisible when you look at each project in isolation.

Dependency Graph

AIMasterTrainer ──feeds knowledge──► CLAUDE.md (global constitution)
     │                                      │
     └──wisdom distillation──────────────────┘
                                            │
                                  all projects benefit

RudraForge ──product ops──► JewelVault / SmartReview / ErrorLens
     │
     └──OPP-018──► rudraforge-graph (indexes RudraForge itself)

Indusynth.com ──parent brand──► SwaPrep (com.indusynth.swaprep)
     │                    └──► SameeraLearn (admin dashboard)
     │                    └──► RudraForge products (privacy policy host)
     │
     └──revenue path──► SwaPrep subscriptions (primary)
                    └──► RudraForge products (secondary)

SwaPrep ──proof point──► Indusynth.com (AI Content Pipeline service)
SameeraLearn ──proof point──► Indusynth.com (AI App Development service)

Convergent Patterns (April 2026)

Pattern 1: Structural enforcement beats documentation governance

Evidence across: AIMasterTrainer (init_guard.py blocks writes), SameeraLearn (content approval gate), SwaPrep (correctAnswer as code comment), RudraForge (pre-release audit hook). Universal principle: Rules operate on willpower. Walls operate on physics. Every project that moved a constraint from a document to a hook stopped violating it. (ML#8, ML#36, ML#63)

Pattern 2: Verification-as-discipline is the dominant pattern emerging

Evidence: AIMasterTrainer/WIS-012 (verification becomes primary discipline when creation is cheap), SwaPrep (scoring gate: 9/9 test cases before release), SameeraLearn (parent approval before child sees), rudraforge-graph (gap detection as primary function). Universal principle: When AI makes generation trivially cheap, the bottleneck shifts permanently to verification. Every project should ask: “Where is our verification layer?” (ML#122, WIS-012 pending)

Pattern 3: Autonomous operation requires dedicated context budget

Evidence: AIMasterTrainer (harvest sessions MUST be dedicated, never mixed), RudraForge (product sessions separate from infrastructure sessions), SwaPrep (content sessions vs app sessions separate). Universal principle: Token throughput for autonomous operation is non-negotiable. Mixing operational contexts degrades both. (ML#121, mastertrainer SKILL.md — “MasterTrainer sessions must be DEDICATED”)

Pattern 4: The intelligence layer compounds across all projects

Evidence: Every project benefits from meta-lessons in CLAUDE.md. WIS-009/010/012 pending integration will apply to session-initializer, SwaPrep pipeline, SameeraLearn review gates, and RudraForge product scoring — simultaneously. Universal principle: Investment in CLAUDE.md and skills pays dividends in every subsequent session across every project.

Active Cross-Project Dependencies

DependencyFromToRisk if broken
yogii@indusynth.com verificationGoogleRudraForge (Play Store)JewelVault launch blocked
Razorpay accountIndusynthSwaPrep (payments)Subscription revenue blocked
Firebase projectIndusynthSwaPrep + SameeraLearnBoth apps offline
Global skills~/.claude/skills/All projectsEvery session degrades
CLAUDE.md~/.claude/All projectsAll behavioral rules lost

Gaps Visible Only From Cross-Project View

  1. No unified monitoring: Each project tracks health separately. No dashboard shows “SwaPrep OK, SameeraLearn 3 P0, AIMasterTrainer 83.8” in one view. (This wiki is the first step.)
  2. Indusynth.com is not commercially active: The parent company website that every project points to for credibility has no paying clients and is password-protected. Revenue risk.
  3. rudraforge-graph could index ALL projects: Currently scoped to RudraForge only. Expanding to index SwaPrep + SameeraLearn session learnings would make this wiki auto-maintained.

Graph View

Backlinks