Jibo-Revival-Group/JiboExperiments
8
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add person-aware favorites and multitenant state scaffolding

Browse Source
This commit is contained in:
Jacob Dubin
2026-05-14 21:15:14 -05:00
parent ec786be797
commit c30363ec9f
5 changed files with 138 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
OpenJibo/docs/feature-backlog.md
View File

@@ -815,6 +815,31 @@ Current release theme:
- at least one legacy prompt pack is running through OpenJibo content instead of hand-authored fallback text
- we have a clear second-wave list for the more conditional MIM families
### 30. Original Personalized Function Inventory
- Status: `discovery`
- Tags: `content`, `docs`, `protocol`
- Why now:
- we are actively porting persona and memory slices, so we need a bounded checklist of the original Jibo charm surfaces
- the goal is to keep the next few passes focused on personality-rich wins instead of letting the work sprawl
- Known sources:
- legacy Jibo OS/Pegasus chitchat and MIM response families
- current OpenJibo persona, memory, and greeting work as the implementation target
- Inventory to track:
- identity and origin questions
- personality and capability questions
- favorite-style prompts like `what is your favorite color`
- mood / affect questions
- follow-up state prompts that should stay warm and locally grounded
- Next pass targets:
- port a small favorites family, starting with `favorite color`
- keep adding 1-3 persona prompts per pass with tests
- prefer source-backed MIM imports when the legacy text is available, and use a temporary runtime reply only when needed to unblock user value
- Exit criteria:
- a stable checklist exists for the original persona surface
- each pass can be scoped to a small batch of prompts
- the backlog makes it obvious what is still missing without losing momentum
## Suggested Order
Before closing `1.0.18`:

19
OpenJibo/docs/release-1.0.19-plan.md
View File

@@ -20,6 +20,25 @@ The goal is to keep compatibility work steady while shipping personality and cap
- start building reusable content hooks for question-vs-command style responses
- keep first implementation rule-based and test-backed
### 1a. Original Personalized Function Inventory
Keep a running checklist of the legacy persona questions and identity surfaces we want to preserve or port:
- identity and origin: `what are you`, `who are you`, `what is Jibo`, `who made you`, `where are you from`
- persona and capability: `do you have a personality`, `what is your job`, `how much do you know`, `what do you want`
- self-description and social charm: `what's your name`, `do you have a nickname`, `do you like being Jibo`, `are there others like you`
- favorite-style prompts: `what is your favorite color`, `what is your favorite food`, `what is your favorite music`
- affect and mood: `how are you`, `are you happy`, `are you sad`, `are you angry`
- conversational follow-ups that should stay local and warm instead of falling into generic chat
Current batch note:
- `favorite color`, `favorite food`, and `favorite music` are the first small favorites-family slice
- the next passes should keep the same pattern and prefer source-backed phrasing whenever the legacy MIM text is available
- if a source-backed legacy line is missing, use a temporary direct reply only to keep the pass moving, then backfill source text later
The goal is to port these in small batches, capture the source-backed phrasing where possible, and keep a test for each batch so the list never becomes a vague backlog graveyard.
### 2. Reliability And Device Proof
- complete update/backup/restore proof path with captures and operator docs