PiNG! Project View — Unified AI Workstation

PiNG! Active

AI-powered desktop workspace — Electron, React, TypeScript, SQLite

ElectronReactTypeScriptSQLiteVite

New Task T

New Note N

Run Workflow

Semantic Search

Chats

56

+3 this week

Agents Active

2

1 running, 1 thinking

Tasks

6/37

16% · 6 overdue

Cost This Month

$14.27

of $50 budget

Live Systems

Active Agents 2 liveManage →

◈

Browser Health C16 Running

Executing gate-check.sh — verifying 1037 tests pass...

opus 4.6

42K tokens

◈

Auto Runner C2 Thinking

Planning AR-C2 implementation based on spec...

opus 4.6

18K tokens

Connections Add →

◉

GitHub

loudbox/ping-skills · 3 open PRs

◈

MCP Servers

3 active · filesystem, browser, preview

▭

Linear

PiNG workspace · synced

♦

Slack

Not connected

Environment

Dev Server

localhost:5173 · vite · 24m

Stop

Electron Main

PID 4821 · electron . · 24m

Stop

Test Watcher

vitest --watch · stopped

Start

Work

Recent Chats All 56 →

Browser Health C162h

UI Talk — Mockups3h

Logger refactor1d

Gate check fixes2d

FTS5 migration3d

Tasks 37Board →

Review PR by EOD46d overdue

Fix MCP hardening C312d overdue

Auto Runner C2-C6in progress

✓

Browser Health — SHIPdone

Calendar Full view →

March 2026

‹

›

Su

Mo

Tu

We

Th

Fr

Sa

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

1

2

3

4

Intelligence

Knowledge 128View all →

Fact

The SDK is a downstream adapter for Gemini wire format

Source conversation · 8:25 AM

Decision

Adopt "Star + Context Horizon" model for memory

UI Talk · yesterday

Workflows 5

⚡

Gate Check & Report

Trigger: on commit

Run

▶

PlanPack Chunk Execute

Manual · chunk + mini-SVQC

Run

Project Memory

System

PING.md — code standards, work modes, gate checks

220 lines · updated 1d ago

Learning

54 learnings (GL-SWEEP, GL-RECEIPT, GL-TEST...)

learnings.md · updated 2d ago

Context Budget14K / 1.0M tokens

Health & Security

Cost & Usage

Opus

Sonnet

1.2M

tokens today

$14.27

this month

$35.73

remaining

Project Health

Tests

5,800+

Type Safety

0 any

File Sizes

Clean

Tasks

16%

Tech Debt

2 items

Audit Log

▶Agent ran gate-check.sh — all gates passed2h ago

✎Edited BrowserManager.ts lines 12-852h ago

✎Created chat-pane-v2.html3h ago

▶Ran npm test — 5,800 tests passed3h ago

▶Git commit: "BH-C16: final chunk SHIP"4h ago

Recent Activity

2h ago

Browser Health C16 — SHIP confirmed. All 16 chunks complete. 1037 browser tests.

3h ago

UI Talk — Chat pane v2 mockup created: chat-pane-v2.html

5h ago

Knowledge — 4 new facts extracted from conversations

1d ago

Tools Wave 2 — SHIP confirmed. 4 chunks complete, 104 new tests.

Folders

💬

All Chats 56

▶

📁

PiNG! Dev 20

Browser Health C16 2h

Auto Runner C2 5h

Logger refactor 1d

Gate check fixes 2d

FTS5 migration 3d

Tools Wave 2 5d

MCP Harden C2 6d

▶

📁

CatSkills 6

Skill Builder 2d

Prompt Templates 4d

Skill Triggers 5d

Agent Routing 1w

Eval Framework 1w

Custom Actions 2w

▶

📁

Memory System 3

Star + Context Horizon 1d

Memory Architecture 3d

Context Budget Design 5d

▶

📁

REPO/Coding 6

Gemini adapter testing 1w

SQLite WAL tuning 1w

IPC refactor 2w

Vite HMR config 2w

ESLint rules 3w

Monorepo structure 3w

▶

📁

Features 3

UI Talk — Mockups 3h

Product Landing Page 1d

Voice Settings 2d

▶

📁

FIXES 4

Crash on large paste 3d

Memory leak sidebar 5d

Dark mode glitch 1w

Scroll jump fix 2w

▶

📁

RAG Tests 1

Embedding benchmark 2w

📁

Unfiled 13

STATS

Total tokens: 4.2M

Total cost: $14.27

Avg per chat: 75K tok

«

Browser Health C16 ★

Executing gate-check.sh — verifying 1037 tests pass...

opus 4.642K tok→ continued from C15

2h ago

284 msgs

UI Talk — Mockups ★

make it uniform across your mockups the workspace list...

opus 4.638K tok

3h ago

96 msgs

Auto Runner C2

Planning AR-C2 implementation based on spec...

opus 4.618K tok→ continued from AR-C1

5h ago

42 msgs

Logger refactor

Finished — structured JSON output added to all services

sonnet 4.68K tok

1d ago

18 msgs

Star + Context Horizon

Adopted Star model for context window management...

opus 4.628K tok

1d ago

64 msgs

Product Landing Page

Hero section with cat mascot animation ready for review

opus 4.615K tok

1d ago

28 msgs

Gate check fixes ★

Fixed console.* violations in 3 files, updated gate-check.sh

opus 4.622K tok

2d ago

56 msgs

Skill Builder

New YAML-based skill definition format working...

opus 4.631K tok

2d ago

72 msgs

Voice Settings

TTS engine selection and pitch controls complete

sonnet 4.69K tok

2d ago

22 msgs

FTS5 migration

Should we use external-content mode or standalone for KB?

opus 4.635K tok

3d ago

128 msgs

Crash on large paste

Fixed — was exceeding IPC message size limit at 5MB

opus 4.614K tok

3d ago

36 msgs

Memory Architecture

Three-tier model: hot (session), warm (project), cold (global)

opus 4.645K tok

3d ago

142 msgs

Prompt Templates

Mustache-style variable interpolation with defaults

sonnet 4.611K tok

4d ago

48 msgs

Tools Wave 2

SHIP confirmed. 4 chunks complete, 104 new tests.

opus 4.662K tok

5d ago

320 msgs

Memory leak sidebar

React effect cleanup was missing on unmount

sonnet 4.66K tok

5d ago

14 msgs

Context Budget Design

Token allocation strategy per memory tier defined

opus 4.619K tok

5d ago

52 msgs

MCP Harden C2 ★

Trust Mechanism: receipts with cryptographic proofs

opus 4.648K tok

6d ago

186 msgs

Gemini adapter testing

thoughtSignature management FROM the event IR

gemini 2.512K tok

1w ago

34 msgs

Embedding benchmark

Comparing voyage-3 vs text-embedding-3-large on our corpus

opus 4.622K tok

2w ago

46 msgs

»

Chat Details

Title

Browser Health C16

Model

opus 4.6

Status

● Running

Tokens

42,180

Messages

284

Created

Mar 9, 2026

Folder

📁 PlanPack Work

Continuation

→ from Browser Health C15

Linked Items

☑ Fix MCP hardening C3 task

📄 BrowserManager.ts file

⚡ Decision: adopt Star + Context... note

Extracted Notes

⚡ Adopt Star + Context Horizon model

⭐ Concurrency via Git Worktrees

Type

All 128

Decision 14

Fact 42

Open Question 8

Context 31

Preference 6

Observation 18

Idea 9

Source

All Sources 128

Auto-extracted 96

Manual 32

Status

Pinned 5

Promoted to Global 12

⚡ Decisionconf: 0.95★

Adopt "Star + Context Horizon" model for memory architecture — central project memory with radiating conversation-level contexts

💬 UI Talk — Mockups·yesterday, 3:42 PM

✎

📌

🔗

↑

🗑

📌 Factconf: 0.92auto-extracted

The SDK is used as a downstream adapter to produce correct Gemini wire format (including thoughtSignature management) FROM the event IR

💬 Source conversation·today, 8:25 AM

✎

📌

🔗

❓ Open Questionconf: 0.78

Should we use FTS5 external-content mode or standalone for the knowledge base?

💬 FTS5 migration·3 days ago

👁 Observationconf: 0.88

Trust Mechanism: "Receipts" with cryptographic proofs could be a differentiator

💬 MCP Harden C2·6 days ago

💡 Ideaconf: 0.82

Stop assuming cross-chat memory means copying full context — share structured knowledge objects instead

💬 UI Talk — Mockups·today

⭐ Preferenceconf: 0.98★

Concurrency should be handled via Git Worktrees — each agent gets its own isolated copy

💬 Browser Health C16·2h ago

Note Details

Type

⚡ Decision

Confidence

0.95 ●

Source

Auto-extracted

Created

Mar 10, 2026, 3:42 PM

Pinned

★ Yes

Linked Items

💬 UI Talk — Mockups chat

☑ Project View UI design task

📄 memory-arch.md file

📁 All Files 18

📁 Unfiled 3

▼📁 Research 5

MD RAG implementation v251.md

TX Phase 4.0 Agentic browser.txt

▼📁 Embeddings 2

MD Vector search notes.md

TX BM25 vs cosine comparison.txt

▼📁 Drafts 6

MD PiNG! Whitepaper.md

MD Memory Recommendations.md

▶📁 UI Mockups 3

▶📁 Final 2

MD Technical White Paper.md

MD Product Brief.md

NameSizeModifiedStatusAI Context

MD

RAG implementation v251.md

Research

111 KB

Jan 2

Draft

📝 Summary

MD

PiNG! Technical White Paper.md

Final

68.8 KB

Dec 19

Canonical

✅ Full

TX

Phase 4.0 Agentic browser.txt

Research

43.7 KB

Jan 5

Draft

📝 Summary

MD

PiNG! Whitepaper.md

Drafts

29 KB

Dec 19

Draft

📝 Summary

TX

the-ping-memory-magic.txt

Unfiled

5.4 KB

Dec 22

Archived

— Excluded

RAG implementation v251.md

111 KB·Markdown·📝 Summary in context

AI Summary

Technical deep-dive on RAG for PiNG!. Covers chunk strategies, embedding models, vector search with SQLite vec extension, and hybrid retrieval. Recommends 70/30 BM25/vector weighting.

Preview

# RAG Implementation v2.51 ## Overview This document outlines the complete RAG pipeline for PiNG!'s knowledge retrieval system... ## 1. Chunking Strategy - Window size: 512 tokens - Overlap: 64 tokens ...

Indexed Files

247

of 842 discovered

Chunks

1,842

avg 380 tokens each

Memories

67

42 project + 25 global

Embeddings

1,842

384-dim vectors

Last Indexed

2h ago

auto-index on

📄 Knowledge Base

🌱 Project Memory

🌐 Global Memory

🌱 AI Context Budget

Token Budget1,484 / 2,000

⚠ 74% used — consider increasing budget

Breakdown

Notes (36 of 545) 846 tok

Files (12) 260 tok

Conversations (2 of 40) 340 tok

Brief 38 tok

BrowserManager.ts

src/electron/services/

Indexed83,240

AgentTaskService.ts

src/electron/services/

Indexed124,860

PING.md

./

Indexed62,480

schema.sql

src/electron/db/

Extracting——

vite.config.ts

./

Excluded——

Type

All 42

Brief 3

Decision 14

Note 25

Origin

Auto 28

User 10

System 4

Briefproject🔧 system

PiNG! is an AI-powered desktop workspace built with Electron, React, TypeScript, and SQLite. Owner: Allan (Loudbox Studios, Philippines). Uses three-role system: Coach 1 (Claude Opus), Coach 2 (ChatGPT), Executor (Claude Code).

Decisionux🤖 auto

Adopt "Star + Context Horizon" model for memory architecture — central project memory with radiating conversation-level contexts

Decisiontechnical🤖 auto

Concurrency handled via Git Worktrees — each agent gets its own isolated copy of the repo

Noteprocess👤 user

Allan prefers PlanPack for multi-file changes and Small Fix mode for quick bug fixes. When in doubt, ask.

Briefproject🔧 system

Current codebase: 5,800+ tests, 0 any types, gate-check.sh enforces quality. Browser Health (16 chunks) and Tools Wave 2 (4 chunks) are SHIPPED.

Notetechnical🤖 auto

Trust Mechanism: "Receipts" with cryptographic proofs could be a differentiator for MCP hardening

Category

All 25

Technical 8

Preferences 5

Identity 3

Process 6

UX 3

Type

Fact 18

Topic Snapshot 7

Factidentity👤 user

Allan is a developer based in the Philippines, running Loudbox Studios. He builds PiNG! as a commercial AI workspace product.

Factpreferences👤 user

Always use bun instead of npm for package management. Never auto-commit — only commit when explicitly asked.

Facttechnical🤖 auto

The SDK is used as a downstream adapter to produce correct Gemini wire format (including thoughtSignature management) FROM the event IR

Topic Snapshotprocess🤖 auto

PlanPack evolution: V12 complete (C26-C39 SHIPPED), MCP-HARDEN in progress (C1-C2 done, C3-C5 pending). 54 learnings recorded. Governance version 2.28.

Facttechnical🤖 auto

better-sqlite3 is synchronous — do NOT use await on DB calls. DI pattern uses get*Service(deps?) / reset*Service() singleton getters.

Topic Snapshotux🤖 auto

Project View UI design: replacing modal-based project interface with full-page workspace. 8 views (Overview, Chats, Notes, Files, Knowledge, Analytics, Tasks, Calendar).

Messages

262

↑ +65% vs last week

Tokens Used

4.2M

↑ +38%

Notes Created

28

↑ 18 auto-extracted

AI Cost

$14.27

of $50 budget

Conversations

56

↑ +3 this week

📊 Daily Activity Messages sent & received

Mon

Tue

Wed

Thu

Fri

Sat

Sun

Sent

Received

💰 Cost by Task Type

Task Type	Cost
Chat Messages	$8.42
File Summaries	$2.14
Note Extraction	$1.86
Smart Search	$0.92
Conv. Summaries	$0.54

📈 Token Usage by Model

Mon

Tue

Wed

Thu

Fri

Sat

Sun

Opus ($11.20)

Sonnet ($2.45)

Haiku ($0.62)

🚀 Background AI Usage

Metric	Value
Background Tasks	142
Cache Hits	68 (48%)
Background Cost	$3.85
Budget Used	28.5%

⚠ Overdue 6

Review PR by EOD

46d overdue

Fix MCP hardening C3

Critical12d

Update learnings registry with Tier-2 fields

5d

Fix the repeating problem

3d

📋 To Do 25+

Auto Runner C2 — implement spec

Mar 15PlanPack

Universal tool calling for all models

Mar 20

Implement project view UI (this mockup!)

Mar 25

+ 22 more tasks...

▶ In Progress 0

No tasks in progress.
Drag a task here to start.

✓ Done 6

Browser Health — SHIP (all 16 chunks)

completed 2h ago

Tools Wave 2 — SHIP

completed 5d ago

🤖 AI Suggestions 3

Tasks extracted from recent conversations:

Implement chat pane v2 layout from mockup

"...make me a mock up" — UI Talk

conf: 0.88

Add Tier-2 fields to learnings registry

"...does not yet have ACTIVE/MUST..." — chat

conf: 0.92may be duplicate

Set up project view as landing page

"...shown first instead of chat..." — UI Talk

conf: 0.85

Sun

Mon

Tue

Wed

Thu

Fri

Sat

1

Sprint Planning

2

PR Review deadline

3

4

5

MCP Harden C3 due

6

7

8

9

BH C16 ✓

10

UI Review

11

🎉 BH SHIP

Standup 9AM

12

13

14

15

AR C2 due

Weekly Sync

16

17

18

Sprint deadline

19

20

Universal tools due

21

22

AR C3-C6 due

23

24

25

Project View UI due

26

27

28

29

30

31

🏁 Month End

1

2

3

4

Sun 8

Mon 9

Tue 10

Wed 11

Thu 12

Fri 13

Sat 14

6 AM

7 AM

8 AM

9 AM

BH C16 ✓

UI Review

Standup 9AM

10 AM

🎉 BH SHIP

11 AM

12 PM

1 PM

2 PM

3 PM

4 PM

5 PM

6 PM

Wednesday, March 11, 2026 Today

6 AM

7 AM

8 AM

9 AM

🔔 Standup 9AM — Reminder

10 AM

🎉 Browser Health SHIP — Milestone (all day)

11 AM

12 PM

1 PM

2 PM

3 PM

4 PM

5 PM

6 PM

7 PM

8 PM

All 8 Prompt 5 Integration 2 Workflow 1

Code Review

Prompt

Structured code review with severity-tiered findings and actionable recommendations

reviewquality

Used 24x

project

Test Generator

Prompt

Generates comprehensive test suites with edge cases and full coverage

testingvitest

Used 18x

project

GitHub PR Bot

Integration

Automated PR creation and review with GitHub API and webhooks

githubpr

Used 22x

project

Full Review Pipeline

Workflow

Chains Code Review → Security Audit → Test Generator

pipelinereview

Used 6x

project

Also available from global library

+ Think Deeper + Security Audit + Refactor Clean + Deploy Checker Browse all 47 catskills →