frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Mobile-MCP: Letting LLMs autonomously discover Android app capabilities

1•weloverust•1h ago
Hi all,

We’ve been thinking about a core limitation in current mobile AI assistants:

Most systems (e.g., Apple Intelligence, Google Assistant–style integrations) rely on predefined schemas and coordinated APIs. Apps must explicitly implement the assistant’s specification. This limits extensibility and makes the ecosystem tightly controlled.

On the other hand, GUI-based agents (e.g., AppAgent, AutoDroid, droidrun) rely on screenshots + accessibility, which gives broad power but weak capability boundaries.

So we built Mobile-MCP, an Android-native realization of the Model Context Protocol (MCP) using the Intent framework.

The key idea:

- Apps declare MCP-style capabilities (with natural-language descriptions) in their manifest.

- An LLM-based assistant can autonomously discover all exposed capabilities on-device via the PackageManager.

- The LLM selects which API to call and generates parameters based on natural language description.

- Invocation happens through standard Android service binding / Intents.

Unlike Apple/Android-style coordinated integrations:

- No predefined action domains.

- No centralized schema per assistant.

- No per-assistant custom integration required.

- Tools can be dynamically added and evolve independently.

The assistant doesn’t need prior knowledge of specific apps — it discovers and reasons over capabilities at runtime.

We’ve built a working prototype + released the spec and demo:

GitHub: https://github.com/system-pclub/mobile-mcp

Spec: https://github.com/system-pclub/mobile-mcp/blob/main/spec/mobile-mcp_spec_v1.md

Demo: https://www.youtube.com/watch?v=Bc2LG3sR1NY&feature=youtu.be

Paper: https://github.com/system-pclub/mobile-mcp/blob/main/paper/mobile_mcp.pdf

Curious what people think:

Is OS-native capability broadcasting + LLM reasoning a more scalable path than fixed assistant schemas or GUI automation?

Would love feedback from folks working on mobile agents, security, MCP tooling, or Android system design.

The Kind of Joke No One Makes Anymore

https://nik.art/the-kind-of-joke-no-one-makes-anymore/
1•herbertl•2m ago•0 comments

Built a Hacker News Client for Terminal

1•shivkmojha•4m ago•0 comments

" AI infrastructure is controlled by companies making toilets, MSG, and glass"

https://twitter.com/Gaurab/status/2024259874276999338
1•ZeljkoS•7m ago•0 comments

Claude Code Skills and 380 agent skills from official dev teams and community

https://github.com/VoltAgent/awesome-agent-skills
1•saikatsg•9m ago•0 comments

Eth.zig: The fastest Ethereum library. Pure Zig. Zero dependencies

https://www.ethzig.org/introduction
2•jedisct1•11m ago•0 comments

AI agents are fast, loose, and out of control, MIT study finds (ZDNET)

https://www.zdnet.com/article/ai-agents-are-out-of-control-mit-study/
1•ildar•11m ago•0 comments

AI Agent Crypto Wallets Create New Legal Risks, Investors Warn

https://www.ccn.com/news/crypto/ai-agent-crypto-wallets-legal-risks-investors-warn/
1•ildar•11m ago•0 comments

Get free Claude max 20x for open-source maintainers

https://claude.com/contact-sales/claude-for-oss
1•zhisme•12m ago•0 comments

Working on Pharo Smalltalk: BPatterns: Rewrite Engine with Smalltalk Style

http://dionisiydk.blogspot.com/2026/02/bpatterns-rewrite-engine-with-smalltalk.html
2•mpweiher•16m ago•0 comments

Show HN: RunVeto – A Simple Kill Switch for Autonomous AI Agents

https://www.runveto.xyz/
1•JDPatel1729•17m ago•0 comments

The LLM App Isn't a Model, It's a System: Designing for Quarterly Model Swaps

https://garybake.com/seams1.html
1•garybake•21m ago•1 comments

An Unbiased OSS Benchmark. For Code Review Agents

https://codereview.withmartian.com
2•alokDT•21m ago•0 comments

The 'Million AI Monkeys' Hypothesis and Real-World Projects

https://ayende.com/blog/203907-B/the-million-ai-monkeys-hypothesis-real-world-projects?key=76196b...
1•ayende•22m ago•0 comments

Merrilin – We built an app to read books

https://tech.stonecharioteer.com/posts/2026/merrilin/
3•stonecharioteer•22m ago•1 comments

Software development now costs less than than the wage of a minimum wage worker

https://ghuntley.com/real/
1•ghuntley•25m ago•0 comments

My personal blog's traffic is 95% AI crawlers this week

https://c0n0.com/posts/ai-traffic/
2•jealousgelatin•27m ago•0 comments

Does journaling help people understand themselves long-term?

2•ashutoshbhatia•28m ago•1 comments

Template for telling the truth on layoffs like at Block

1•nutanc•28m ago•0 comments

Reddit disables any access to R/all on mobile

https://www.reddit.com/r/help/s/vnPnnrsUPm
1•embedding-shape•29m ago•0 comments

HikmaAI – The AI Agent Supply Chain is Broken. Here is how we fix it

https://twitter.com/HikmaAi_/status/2027286845290098768
2•NibrasSB•29m ago•0 comments

How to Run Services on a Linux Server

https://huijzer.xyz/posts/150/how-to-run-services-on-a-linux-server
1•huijzer•31m ago•1 comments

AI voice agents for hotels: lessons from 15,910 real guest calls

https://polydom.ai/blog/what-16-000-hotel-phone-calls-taught-us-about-ai-voice-agents
2•wastemaster•32m ago•1 comments

Show HN: I Created an Interactive Resume Space Invader Game

https://breezko.dev
3•breezk0•32m ago•0 comments

Show HN: Asupersync, the Cancel-Correct Async Runtime for Rust

https://asupersync.com/
2•eigenvalue•32m ago•0 comments

Atomic GraphRAG Explained: The Case for a Single-Query Pipeline

https://memgraph.com/blog/atomic-graphrag-explained-single-query-pipeline
2•mbuda•32m ago•1 comments

The Teaching Method That Can't Fail

https://substack.com/home/post/p-187807163
1•artbristol•32m ago•0 comments

AI Assisted Coding

https://kirit.com/Blog:/2026/02-27-AI%20Assisted%20coding
1•KayEss•34m ago•0 comments

Apple announces F1 details, and a surprising Netflix partnership

https://sixcolors.com/post/2026/02/apple-announces-f1-details-and-a-surprising-netflix-partnership/
1•tosh•34m ago•0 comments

Addressing AI-slop in security reports

https://github.com/apache/logging-log4j2/discussions/4052
1•Tomte•34m ago•0 comments

I swarms can threaten democracy by manufacturing fake public consensus

https://techxplore.com/news/2026-01-generation-disinformation-ai-swarms-threaten.html
1•xyzal•38m ago•0 comments