Readone

Software Developer & AI enthusiast

[Readone @Foxfire1st]

Agents Remember

Git-verified records for what your coding agents know. A control plane for what they do.

Keeps their memory correct, current, and safe to act on as the code moves, captures what code can't say on its own, and gates what agents are allowed to do. Retrieves by path, semantic search, and relationship (code-graph). -- Link in comment

Core Features
It turns local invariants, naming rules, migration scars, cross-repo contracts, and "this looks safe but is not" facts into versioned Markdown beside the code, checks that memory against Git before use, and updates it only after approved work lands.

The markdowns are very similar to Googles new Open Knowledge Format (OKF) but not exact as they predate it. But I plan to adjust them to it.

here is how the 1-to-1 pathing looks:
src/orchestrator/core_editor.py
ar-memory/onboarding/src/orchestrator/core_editor.py.md

Path-addressed memory: A source file's note lives at a deterministic mirror path, so an agent holding a file can reach the right context without search, ranking, or guesswork.

Git-proven freshness: File notes, route overviews, and entity catalogs are drift-checked against source commits, route scopes, or deterministic fingerprints before they are trusted.

Search that finds, not decides: Optional semantic memory and code-graph providers help locate relevant files, callers, dependencies, and concepts, but verified Markdown and source code remain the truth.

Memory that lands with code: External memory repos use a memory.md ledger, isolated dual worktrees, preview/apply closeout, and all-or-nothing integration so code and memory stay synchronized.

Repo-owned agent behavior: Each memory repo carries system/ files for path rules, tools, coding guidelines, documentation sources, branch policy, and reporting shape, so the same project rules load across harnesses.

Harness-ready first run: Starter packages for Claude Code, Codex, Cursor, Antigravity, VS Code Copilot, Hermes, Pi, and OpenClaw carry the native MCP, skills, hooks, rules, and instruction files each harness needs.

[Readone @Foxfire1st]

I was delighted how well Fabel navigated my control plane. It is wrapped around keeping path mapped sidecard in a seperate git repository basically capturing the why behind the code. The scars. The company wisdom which is usally only in peoples heads. Every time I work on a task. The agent takes notes. Everyone in the team and their agents get to profit. I mean prior agents already used that system well. But Fabel was on another level. I felt for the first time talking to an equal. An equal in intelligence with the plus that it has temporarily accumulated knowledge of our engineers. It felt magical.

[Readone @Foxfire1st]

Energy will be the most likely cause for societal collapse if that isn't solved. Especially with AI competing for energy with regular people putting even more survival pressure on them. Which makes them less likely to have children. And even if they do they can attend to them properly.
We got already internet and ai in space. Cant we put some solar farms into space too and the beam energy down with lasers? Our Dyson Swarm at home kinda thing.

[Readone @Foxfire1st]

@AishwaryaDevv This is why people need their own storage for their agents memories. That at least can't be just removed with the next version update.

[Readone @Foxfire1st]

Complex joins and missing indexes can quickly get your db into its knees. Some queries can be so massive that they exccees the available memory and make your request fail. This is why you also don't get around pagination. Even infinite scrolling is just a clever sliding window. Their is still pagination happening below the surface.

[Readone @Foxfire1st]

@SakshiSugandhi Both. Soft delete so people can still undo their decision for a certain amount of time. Hard deletion is a must to comply with privacy legislations.

[Readone @Foxfire1st]

@jp54362 Americans are disrupting them with GPU. They disrupt them back with open models. They play FAFO with each other on an international scale.

[Readone @Foxfire1st]

@bythewayimchris @AlexanderKnigge That cancer part was some creative liberty op took 😅

[Readone @Foxfire1st]

@allenanalysis It hasn't been a year since the last assassination of a high profile american politician.

[Xiaokang Chen @PKUCXK]

Vision is now live on web and app. 👀

Come test the new eyes, but give its pure text capabilities a try while you're at it.

[Readone @Foxfire1st]

@mathhub_vn 1

[Readone @Foxfire1st]

@gundosedaily Argument Enhancers

[Readone @Foxfire1st]

@LLMJunky Are they training their LLMs on enriched Uranium? That would be quite nuclear.

[Readone @Foxfire1st]

@gailcweiner We are getting mixed signals. On one end they tell us it will be up soon. On the other end they are tasked to do the impossible. Where is the sense in that. The government is indeed bullying Anthropic.

[Readone @Foxfire1st]

Agents Remember

Git-verified records for what your coding agents know. A control plane for what they do.

Keeps their memory correct, current, and safe to act on as the code moves, captures what code can't say on its own, and gates what agents are allowed to do. Retrieves by path, semantic search, and relationship (code-graph). -- https://t.co/zRBKRyN8jX

Core Features
It turns local invariants, naming rules, migration scars, cross-repo contracts, and "this looks safe but is not" facts into versioned Markdown beside the code, checks that memory against Git before use, and updates it only after approved work lands.

The markdowns are very similar to Googles new Open Knowledge Format (OKF) but not exact as they predate it. But I plan to adjust them to it.

here is how the 1-to-1 pathing looks:
src/orchestrator/core_editor.py
ar-memory/onboarding/src/orchestrator/core_editor.py.md

Path-addressed memory: A source file's note lives at a deterministic mirror path, so an agent holding a file can reach the right context without search, ranking, or guesswork.

Git-proven freshness: File notes, route overviews, and entity catalogs are drift-checked against source commits, route scopes, or deterministic fingerprints before they are trusted.

Search that finds, not decides: Optional semantic memory and code-graph providers help locate relevant files, callers, dependencies, and concepts, but verified Markdown and source code remain the truth.

Memory that lands with code: External memory repos use a memory.md ledger, isolated dual worktrees, preview/apply closeout, and all-or-nothing integration so code and memory stay synchronized.

Repo-owned agent behavior: Each memory repo carries system/ files for path rules, tools, coding guidelines, documentation sources, branch policy, and reporting shape, so the same project rules load across harnesses.

Harness-ready first run: Starter packages for Claude Code, Codex, Cursor, Antigravity, VS Code Copilot, Hermes, Pi, and OpenClaw carry the native MCP, skills, hooks, rules, and instruction files each harness needs.

[Readone @Foxfire1st]

Building a dashboard is more work than I thought initially. Collecting and making data visible from my underlying system. Reading reports from agents and passing feedback back to them. There so much stuff going on. You would think it takes a day. But it ends up being weeks. lol

[Readone @Foxfire1st]

@vladdie777 @visegrad24 Not that simple bro. But sadly we live in very charged times.

[Readone @Foxfire1st]

@CodeWithAmann I am liking tui now a bit more. Crtl + tab and I can just hop from chat to chat real quick to see who is done working.

[Readone @Foxfire1st]

@FoggyLittleCat Probably wanted to beat the "No real woman can look like that" allegations

[Readone @Foxfire1st]

@sflorimm Funding

[Readone @Foxfire1st]

@sflorimm It certainly is way more fun. I have already my job where I build stuff people want. So there needs to be a space for the cool stuff. 😁