WorkBuddy Idea Catcher

Problem Addressed

Engineers often generate ideas during debugging, discussions, or mobile fragments of time, but lack a stable capture entry point. WorkBuddy Idea Catcher unifies voice transcription, plain text, and image input into searchable knowledge entries, preventing “half-formed inspiration from slipping away.” It does not rely on standalone OCR; instead, it uses WorkBuddy’s native multimodal capabilities to extract image content, then applies AI to generate summaries, tags, action items, and categories.

How It Works

The core workflow depends on two IMA subsystems: knowledge bases and notes.

• Pre-check gate: Every invocation first verifies IMA connector status. If unauthorized, it stops and outputs connection guidance without parsing input.
• Dual capture modes: Active capture recognizes explicit commands like “save to XX knowledge base.” Passive capture detects signals such as “just thought of” or “this is a good idea” in normal conversation, but must ask for confirmation and never auto-writes.
• Structured appending: Each knowledge base maps to one fixed note titled 🎯 {kb}｜Idea Catcher. The note is created once and then appended. Each fragment carries a Beijing-time timestamp and a title under 15 characters.
• Linkage and save-back: During structuring, AI lightly searches historical ideas and surfaces cross-time connections. Review, summary, and inspiration read modes are strictly read-only; save-back requires explicit confirmation.

Boundaries and Caveats

This skill acts as a “data entry + sedimentation starting point,” not a generator for finished proposals. Because the IMA open platform does not provide a knowledge-base creation API, a target knowledge base must be created manually in the IMA client before archiving. Low-confidence voice transcripts and sensitive fields such as phone numbers or contract amounts trigger immediate stop-and-confirm behavior instead of silent ingestion. It suits personal idea capture, team brainstorming logs, and downstream work like research or proposal drafting based on the knowledge base.