Quick poll: what are you trying to do with this repo? (and what should I build next)

[ballred]

Hey folks — I’m trying to figure out what real job you want this repo to do so I can prioritize the right next steps.

If you only do one thing: vote in the poll.
If you have 60–90 seconds: drop a comment using the template below.

I’ll summarize results publicly and use them to pick the next milestone. Thanks!

[ballred]

60‑second comment template (optional, but extremely helpful). Copy/paste and fill what you can:

1. Where are you right now? (pick one)

• Haven’t tried it yet
• Tried briefly (≤30 min)
• Set it up and use it sometimes
• Using it actively (daily/weekly)
• Forked and customizing

2. What outcome are you hoping for?
   Example: “A reliable weekly review that keeps projects moving” / “A repeatable agent workflow I can adapt to my own repo”

3. Biggest blockers

• Setup friction / too many steps
• Unclear first “win” after installing
• I don’t want my notes touched automatically (safety / control)
• Privacy / data leaves my machine concerns
• Cost / access to Claude tooling
• I’m primarily mobile and this feels desktop‑only
• I already have a vault structure; I want only the automation
• Other: ___

4. What would make it 10× better for you?

• A “guided install” that gets to a first successful workflow fast
• A minimal Obsidian plugin UI (run workflows from a command palette, write back into notes)
• A “maintenance pack” of agents (weekly review, inbox triage, tag/link cleanup, project nudges)
• A “bring your own vault” mode (automation without adopting the structure)
• Better safety rails (diff previews, PR-style changes, allowlist of editable files)
• Example vaults / walkthrough videos / before-after demos
• Integrations (web clipper, Readwise-like ingestion, voice memos, email/calendar)
• Other: ___

5. If you did try it: what was the first thing you wanted it to do? (One concrete workflow request is gold.)

[lambda-sf]

60‑second comment template (optional, but extremely helpful). Copy/paste and fill what you can:

 1. Where are you right now? (pick one)


 * [ x ]  Haven’t tried it yet

 * [ ]  Tried briefly (≤30 min)

 * [ ]  Set it up and use it sometimes

 * [ ]  Using it actively (daily/weekly)

 * [ ]  Forked and customizing


 2. What outcome are you hoping for?
    Example: “A reliable weekly review that keeps projects moving” / “A repeatable agent workflow I can adapt to my own repo”

 3. Biggest blockers


 * [ ]  Setup friction / too many steps

 * [ ]  Unclear first “win” after installing

 * [ ]  I don’t want my notes touched automatically (safety / control)

 * [ ]  Privacy / data leaves my machine concerns

 * [ ]  Cost / access to Claude tooling

 * [ ]  I’m primarily mobile and this feels desktop‑only

 * [ x ]  I already have a vault structure; I want only the automation

 * [ ]  Other: ___


 4. What would make it 10× better for you?


 * [ ]  A “guided install” that gets to a first successful workflow fast

 * [ x ]  A minimal Obsidian plugin UI (run workflows from a command palette, write back into notes)

 * [ ]  A “maintenance pack” of agents (weekly review, inbox triage, tag/link cleanup, project nudges)

 * [x]  A “bring your own vault” mode (automation without adopting the structure)

 * [x]  Better safety rails (diff previews, PR-style changes, allowlist of editable files)

 * [ ]  Example vaults / walkthrough videos / before-after demos

 * [x]  Integrations (web clipper, Readwise-like ingestion, voice memos, email/calendar)

 * [ ]  Other: ___

As soon as I saw Obsidian CLI I was hoping someone would pull something off like this. I use obsidian, markdown, and claude with tool usage heavily.

• a BYOV setup would be perfect, with some way to bake in some context that scaffolds the agent/prompt so it doesn't forget to follow the vault structure.
• integration with pandoc would mean better flexibility with document conversion.

this + supernote for handwritten notes + hooks, tools, and "cron" tasks + a general purpose skill routing system would be so awesome. I have a prototypical kludge right now and it's been amazing how quickly I can teach myself new problems by generating problem sets from notes and the right prompt and validation system, and I want to extend it to every day usage/PM work. the biggest nuisance has been a lack of crontasks, but I can't set those up on a corporate box.

[ballred]

@lambda-sf Thanks for this — really glad the vision resonates, especially the BYOV piece.

/adopt just shipped. It's a 6-phase skill that scans your existing vault structure, detects your organization method (PARA, Zettelkasten, LYT, flat, custom), maps your folders to PKM roles interactively, and generates all the config files scoped to your actual directory names. Your vault structure stays untouched — it only adds .claude/ scaffolding and a root CLAUDE.md that teaches the agent your layout. All the hooks (session-init, auto-commit, statusline) now read env vars with fallback defaults, so everything Just Works whether you started from the template or adopted an existing vault.

On the other requests:

• Pandoc integration — interesting angle. What's your main conversion flow? Markdown → PDF for sharing, or ingesting non-markdown sources into the vault?
• Cron tasks — I hear you on corporate boxes. Claude Code hooks fire on session events (start, file write, etc.) which covers some of the "periodic check" use cases. For true scheduled runs, the constraint is Claude Code being a CLI tool — it needs an active session. Would a /review that's smart enough to catch up on missed days scratch part of that itch?
• Problem set generation from notes — that's a killer use case. The skill system is extensible (just a SKILL.md file with instructions) so building a /study or /practice skill that reads your notes and generates problem sets would be straightforward to add. Would love to hear more about what your prototype does.

What's your vault structure look like? Curious whether /adopt handles your layout cleanly or if there are edge cases to iron out.