JARVIS
Use this page when you need to understand how Helpifyr governs shared skill, handoff, trace, and learning families without turning that governance into a second execution system.
When to use this page
- You need to know why shared families such as skill lifecycle, trace envelopes, or learning signals do not drift into repo-local semantics.
- You are checking whether a lifecycle, promotion, or learning surface is draft, active-limited, active, or blocked.
- You need the public-safe explanation of governed skill lifecycle before reading product-specific learning or execution detail.
Prerequisites
- You have read Agent Capability Plane.
- You want the governance layer that sits above individual skill packs or learning consumers.
Architecture / Flow
Step-by-step procedure
1. Treat JARVIS as governance truth, not as a second runtime
JARVIS anchors one canonical lifecycle and family vocabulary for shared skill, handoff, trace, and learning work.
That means:
- Fabric owns the governance truth
- downstream repos consume it read-only
- local execution or runtime behavior is still owned elsewhere
2. Keep lifecycle and promotion posture explicit
JARVIS exists so repos do not invent local meanings for:
- draft versus active-limited versus active posture
- family admission and visibility
- promotion-aware learning or skill lifecycle work
- closure and resolver posture across repos
3. Use family truth before you infer a governed lane
Shared families such as skill-lifecycle.v1, skill-promotion.v1, trace-envelope.v1, or learning-signal.v1 should be read from the canonical JARVIS surfaces first.
This keeps:
- producer and consumer expectations aligned
- learning posture reviewable
- fail-closed behavior explicit when a family is still draft or blocked
4. Keep execution and adoption separate from governance visibility
A family being visible in JARVIS does not mean every downstream runtime has already implemented or enforced it.
Public-safe reading should distinguish:
- governance truth
- downstream adoption evidence
- runtime execution ownership
Verification
This page is being applied correctly when a reader can:
- explain why JARVIS extends Fabric governance instead of becoming a second execution system
- distinguish family visibility from downstream implementation or enforcement
- explain why missing or conflicting governed posture must stay fail-closed
Common failure modes
Treating JARVIS as a runtime product
Problem:
- readers assume JARVIS is the place where work executes rather than the place where family and lifecycle semantics are governed.
Better path:
- treat JARVIS as governance truth, then follow the downstream owner repo for actual runtime behavior
Treating visible family registration as complete downstream adoption
Problem:
- a shared family is visible in registry truth and that visibility gets mistaken for finished runtime rollout.
Better path:
- read governance truth together with downstream adoption or readiness posture before calling a lane complete