Hacker News: Bibabomas

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 06:41:09 +0000

What do you mean by this exactly?

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 06:29:37 +0000

Hey, we do a couple of things specifically to prevent supply-chain attacks. We use trusted publishing on PyPI, and --exclude newer for uv's package resolution. We also try to use the least amount of dependencies possible. A transitive dependency could in theory still be problematic though, e.g. if there's a supply-chain attack on numpy.

The tool itself is fully local though, so there's no real security risks there, there are no outbound network calls or anything like that.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 06:22:54 +0000

Yeah this is a good point, but it's complicated to do well since semble then has to be aware of everything else that happens in a session, which would then make it more intrusive (and it's deliberately designed as a local, non-intrusive alternative to other solutions). I'm thinking that perhaps we can do an isolated opt-in benchmark for this perhaps.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 06:20:49 +0000

In practice this rarely happens though, at least in practice I rarely see agents "grep -C N" or something like that on files it didn't read yet. I use Claude Code and OpenCode extensively, and especially during the first pass through a codebase that is not well understood the agent often just does "cat file" or something similar and gets the entire file in context first, and only then starts doing more finegrained searches, but at that point you already have a lot of irrelevant context in memory. I think the whole value proposition of semble is that you don't have to do that initial read at all and can instead get the right (small) context bits. If you experience is different, would you mind sharing what your setup is like, e.g. how do you get the agent to read less?

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 05:16:23 +0000

I've had the same experience with RTK, where my agent got stuck in a loop with a faulty RTK command and could not escape it since RTK hard overwrites anything automatically. I've uninstalled it again for the time being.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 05:14:39 +0000

Nice! Let us know if you have any feedback or results to share, would be happy to do the same.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 05:12:11 +0000

Sorry to hear about the MCP integration, that's definitely something we'll look into. If you have any info about your system or how to reproduce it please let me know. Very nice to hear about the results, thanks for checking this! The variance is interesting to see, that's probably non-determinism in the LLM rather than semble since semble is deterministic. But I'm guessing we can make that better with the prompt, I'll look into this.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 05:09:48 +0000

The first is hard to test for us unfortunately since we don't use Cursor. But the Claude thing is interesting. I think that providers (especially the ones that directly sell LLM calls like Anthropic) are not incentivised per se to think about token efficiency vs performance, so if you're chasing pure performance, just loading the full codebase into memory might still be the "benchmark topping" way to go. I think the dust hasn't really settled yet and we'll likely see a lot of changes in the coming year about what's the "correct" way to solve it. It might be different based on your harness/budget/model as well.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 05:02:14 +0000

Serena does a lot more than semble (I actually used serena before building this and didn't like how much it does by default). That also made it hard to see if it was actually working well with how many moving parts there are. Semble only does 1 thing: very quick code search, that's it. Context-mode I have not used before though, I will have a look at that, thanks for sharing!

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 04:58:58 +0000

Afaik many harnesses ship the "default" which is grep+read (like Claude Code). But I agree, IMO it's a weird gap. To be fair I don't think providers are that incentivised to reduce token burn at the moment, but my guess is that that will change and tools like this will become at least an natively supported option in some harnesses.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 04:57:27 +0000

Interesting, will have a look at this, thanks for sharing.

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 04:52:03 +0000

Hey, this is something we're actively working on, but this is hard (and expensive) to do well across harnesses/models. The grep pretraining thing is very interesting though, I've noticed the same. E.g. Sonnet 4.6 seems to trust semble but Opus 4.7 less so. I'm hoping we can quantitatively test this and improve it when we have proper benchmarks for this as well. If you do have any feedback though let me know!

New comment by Bibabomas in "Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep"

Bibabomas — Mon, 18 May 2026 04:44:50 +0000

Hey, codebase-memory-mcp and semble are not exactly the same, but it's an interesting comparison, I'll put it on the todolist to check that out and add it to our benchmarks if feasible. If you ever get a chance to do this same comparison with semble it would be super useful feedback since these "real" scenarios are hard to benchmark/replicate.

Show HN: Semble – Code search for agents that uses 98% fewer tokens than grep

Bibabomas — Sun, 17 May 2026 15:37:07 +0000

Hey HN! We (Stephan and Thomas) recently open-sourced Semble. We kept running into the same problem while using Claude Code on large codebases: when the agent can't find something directly, it falls back to grep, reading full files or launching subagents. This uses a lot of tokens, and often still misses the relevant code. There are existing tools for this, but they were either too slow to index on demand, needed API keys, or had poor retrieval quality.

Semble is our solution for this. It combines static Model2Vec embeddings (using our latest static model: potion-code-16M) with BM25, fused via RRF and reranked with code-aware signals. Everything runs on CPU since there's no transformers involved. On our benchmark of ~1250 query/document pairs across 63 repos and 19 languages, it uses 98% fewer tokens than grep+read and reaches 99% of the retrieval quality of a 137M-parameter code-trained transformer, while being ~200x faster.

Main features:

- Token-efficient: 98% fewer tokens than grep+read

- Fast: ~250ms to index a typical repo on our benchmark, ~1.5ms per query on CPU (very large repos may take longer)

- Accurate: 0.854 NDCG@10, 99% of the best transformer setup we tested

- MCP server: drop-in for Claude Code, Cursor, Codex, OpenCode

- Zero config: no API keys, no GPU, no external services

Install in Claude Code with: claude mcp add semble -s user -- uvx --from "semble[mcp]" semble

Or check our README for other installation instructions, benchmarks, and methodology:

Semble: https://github.com/MinishLab/semble

Benchmarks: https://github.com/MinishLab/semble/tree/main/benchmarks

Model: https://huggingface.co/minishlab/potion-code-16M

Let us know if you have any feedback or questions!

Comments URL: https://news.ycombinator.com/item?id=48169874

Points: 303

# Comments: 104

Show HN: Agentcheck – Check what an AI agent can access before you run it

Bibabomas — Sun, 08 Mar 2026 16:05:43 +0000

Hey HN! I've just open-sourced agentcheck, a fast, read-only CLI tool that scans your shell and reports what an AI agent could access: cloud IAM credentials, API keys, Kubernetes contexts, local tools, and more.

Main features:

- Broad coverage: scans AWS, GCP, Azure, 100+ API key environment variables and credential files, Kubernetes, Docker, SSH keys, Terraform configs, and .env files

- Severity levels: every finding is tagged LOW, MODERATE, HIGH, or CRITICAL so you know what actually matters

- CI/CD integration: run agentcheck --ci to fail a pipeline if findings exceed a configurable threshold, with JSON and Markdown output for automation

- Configurable: extend it with your own env vars, credential files, and CLI tool checks via a config file

When you hand a shell to an AI agent, it inherits everything in that environment: cloud credentials, API keys, SSH keys, kubectl contexts. That's often more access than you'd consciously grant, and it’s hard to keep track of what permissions your user account actually has. Agentcheck makes that surface area visible before you run the agent.

It’s a single Go binary, no dependencies. Install with Homebrew:

brew install Pringled/tap/agentcheck

Code: github.com/Pringled/agentcheck

Let me know if you have any feedback!

Comments URL: https://news.ycombinator.com/item?id=47298399

Points: 4

# Comments: 1