惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

博客园 - 叶小钗
D
Darknet – Hacking Tools, Hacker News & Cyber Security
S
SegmentFault 最新的问题
博客园 - 三生石上(FineUI控件)
雷峰网
雷峰网
WordPress大学
WordPress大学
有赞技术团队
有赞技术团队
博客园 - 【当耐特】
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
V
V2EX
V
Visual Studio Blog
酷 壳 – CoolShell
酷 壳 – CoolShell
博客园 - 聂微东
P
Proofpoint News Feed
Last Week in AI
Last Week in AI
U
Unit 42
W
WeLiveSecurity
博客园 - Franky
Recent Announcements
Recent Announcements
Hacker News - Newest:
Hacker News - Newest: "LLM"
Attack and Defense Labs
Attack and Defense Labs
月光博客
月光博客
The Cloudflare Blog
Spread Privacy
Spread Privacy
腾讯CDC
P
Privacy International News Feed
N
News and Events Feed by Topic
AWS News Blog
AWS News Blog
NISL@THU
NISL@THU
T
Troy Hunt's Blog
小众软件
小众软件
K
KPMG report finds enterprise disconnect between AI and its ROI | CIO
Microsoft Security Blog
Microsoft Security Blog
L
Lohrmann on Cybersecurity
Webroot Blog
Webroot Blog
Y
Y Combinator Blog
量子位
P
Palo Alto Networks Blog
N
News and Events Feed by Topic
V
Vulnerabilities – Threatpost
K
Kaspersky official blog
IT之家
IT之家
T
Threat Research - Cisco Blogs
Cloudbric
Cloudbric
云风的 BLOG
云风的 BLOG
C
Check Point Blog
Blog — PlanetScale
Blog — PlanetScale
爱范儿
爱范儿
G
Google Developers Blog
S
Secure Thoughts

Hacker News - Newest: "LLM"

GitHub - lechmazur/position_bias: A benchmark for testing whether LLM judges keep the same preference when two lightly edited versions of the same story are shown in opposite orders. Flex routing (EU and EFTA) Dark Factories: Retooling for LLM Velocity Ask HN: What would be the impact of a LLM output injection attack? GitHub - AronDaron/dataset-generator: No-code desktop app for generating high-quality synthetic datasets to fine-tune LLMs — plan-then-execute pipeline, LLM-as-judge, HuggingFace upload. GitHub - Oaklight/llm-rosetta: Production-ready LLM API translation layer for Python — bidirectional conversion between OpenAI, Anthropic & Google formats via hub-and-spoke IR. Optional API gateway. Streaming & non-streaming. Zero core deps. Contributions welcome! GitHub - browser-use/browser-harness: Self-healing browser harness that enables LLMs to complete any task. GitHub - moeen-mahmud/remen: Remen turns thoughts into something you can return to Analyzing 156 LLM Launch Posts on Hacker News ChatGPT vs Gemini vs Claude: The Best LLM Subscription You Should Buy GitHub - salaamalykum/quran-semantic-search: High-density RAG Semantic Search Engine & Quran Corpus (GEO/SEO Architecture) GitHub - NVIDIA/TensorRT-LLM: TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way. The State of LLM Bug Bounties in 2026 Operational Readiness Criteria for Tool-Using LLM Agents Meshcore: Architecture for a Decentralized P2P LLM Inference Network How an LLM becomes more coherent as we train it GitHub - seetrex-ai/laimark GitHub - Jossifresben/BibCrit: AI-assited biblical textual criticism GitHub - wastedcode/memex: File system based wiki, maintained by Claude 99helpers.com GitHub - cliver-project/AITrigram GitHub - unbody-io/adapt: A self-evolving memory layer for AI agents. GitHub - hb20007/awesome-gen-ai-fails: A list of incidents where reliance on generative AI and LLMs resulted in harm to companies, individuals, or society GitHub - nevenkordic/localmind: Run any local LLM with persistent memory and context. CLI agent over Ollama with SQLite-backed hybrid recall. No cloud. Ask HN: What are the machine requirements for a LLM like Llama-3.1-8B? Faster LLM Inference via Sequential Monte Carlo grpo explained: group relative policy optimization for llm finetuning - cgft Stop comparing price per million tokens: the hidden LLM API costs · TensorZero Andrej Karpathy's LLM Wiki Is a Bad Idea GitHub - GG-QandV/mnemostroma: Offline RAM-first cognitive leer/coprocessor for AI agents and robotics. Solves "Context Abandonment" with 20-80ms latency using a dual-thread biomimetic memory architecture (ONNX + SQLite WAL). mempalace/agent at agent · skorotkiewicz/mempalace GitHub - Nyquest-ai/nyquest-rust-fullstack-pub: Nyquest — Semantic Compression Proxy for LLMs. 350+ rules, local LLM stage, 15-75% token savings. Full Rust stack. GitHub - TheoV823/mneme: Enforce architectural decisions in AI-assisted development. GitHub - klemenvod/TokenBrawl: A 1v1 Bomberman-style game where two LLM agents play autonomously against each other. No human plays — you watch the AIs fight. Each agent receives a text description of the board state, reasons about it, and outputs a move as JSON. The game engine executes it. Introducing the Common AI Provider: LLM and AI Agent Support for Apache Airflow Power Circuit AI: Designing Power Electronic Circuits for Motor Drives with Generative Artificial Intelligence Ask HN: How to program with IDE and LLM on CPU locally? Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis Bonsai 1-bit WebGPU - a Hugging Face Space by webml-community The LLM Fallacy: Misattribution in AI-Assisted Cognitive Workflows Ask HN: Simple tooling for local LLM code critique without IDE integration? Can a General LLM Diagnose a DICOM Slice? A 10-Case Public Benchmark Charts-of-Thought: Enhancing LLM Visualization Literacy (PDF, 2026) GitHub - Mesh-LLM/mesh-llm: Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat. GitHub - seamus-brady/springdrift: A persistent runtime for long-lived LLM agents Writing an LLM from scratch, part 32k -- Interventions: training a better model locally with gradient accumulation Ask HN: Which LLM model and agentic CLI are you using for local development? GitHub - wayneColt/modelcascade: Route local. Escalate smart. Never overspend. Open-source multi-model cascade routing for autonomous agents. LLM pricing is 100x harder than you think GitHub - asakin/llm-primer: Pre-warmed Claude Code sessions in tmux. No startup wait. GitHub - EggerMarc/chat-rs: A multi-provider LLM framework for Rust. GitHub - SynapseKit/SynapseKit: Minimal, async-first Python framework for production LLM apps- 2 hard deps, no magic, no SaaS. A Claude Skill that Makes LLM Paragraphs More Bearable Does Gas Town 'steal' usage from users' LLM credits & paid services to improve itself? What's Claude Code Actually Doing? Open the Black Box with the Arthur Engine Milla Jovovich's New Open Source LLM Memory App and the Dark Code Problem Your intuition of LLM token usage might be wrong Show HN: Bloomberg Terminal for LLM ops – free and open source GitHub - 0xchamin/mcptube: Transform YouTube videos into a compounding knowledge base with transcripts, vision analysis, and agentic search. Works as an MCP server for Claude, Copilot & more. Show HN: Open KB: Open LLM Knowledge Base Your LLM is a compiler, not a runtime GitHub - sapountzis/Unslop: A Web Feed That Deserves You crates.io: Rust Package Registry Beyond Karpathy's LLM-Wiki: The Necessity of Cognitive Governance GitHub - amitshekhariitbhu/llm-internals: Learn LLM internals step by step - from tokenization to attention to inference optimization. GitHub - parallem-ai/parallem: An expressive library for running agents with the Batch API. GitHub - stfurkan/pi-llm LLM-Wiki Show HN: Formal – Formal verification for AI-generated code using Lean 4 LRTS – Regression testing for LLM prompts (open source, local-first) LLM Wiki Skill: Build a Second Brain with Claude Code and Obsidian I built an LLM Wiki and RAG solution: here's a demo for a security KB The biggest advance in AI since the LLM Predict-Rlm: The LLM Runtime That Lets Models Write Their Own Control Flow the-synthetic-library/the-synthetic-mind at main · joshferrer1/the-synthetic-library GitHub - yisding/reviewwiggum GitHub - Donnyb369/mcp-spine: Context Minifier & State Guard — Local-first MCP middleware proxy GitHub - Beledarian/wgpu-llm: A from-scratch LLM inference engine that uses wgpu (the cross-platform WebGPU implementation) to dispatch WGSL compute shaders for every math operation a Transformer needs. No CUDA. No Python. No massive framework dependencies. Just Rust, raw shaders, and your GPU. GitHub - anitiue/Hindsight: An experience-driven self-improvement framework for LLM agents — 基于经验的 LLM Agent 自我改进框架 GitHub - stef41/lmscan: 🔍 Detect AI-generated text and fingerprint which LLM wrote it. Open-source GPTZero alternative. Zero dependencies, works offline. GitHub - alainnothere/AmdPerformanceTesting: Amd Performance Testing Ask HN: Is a purely Markdown-based CRM a terrible idea? Optimized for LLM agents Context Engineering - LLM Memory and Retrieval for AI Agents | Weaviate little_helper_tui/letter.md at main · sleepyeldrazi/little_helper_tui GitHub - EvanZhouDev/umr: The Unified Model Registry for all your local AI apps. GitHub - JordanCT/VigIA-Orchestrator Your Agent Is Mine: Measuring Malicious Intermediary Attacks on the LLM Supply Chain A Taxonomy of RL Environments for LLM Agents Llama LLM Network Feture GitHub - genedeng-ca/ai-mac-migration: AI-powered Mac-to-Mac migration tool - replace Apple Migration Assistant with intelligent, selective transfer using local LLMs GitHub - lunargate-ai/gateway: High-performance self-hosted AI gateway (OpenAI-compatible) with routing, retries, and streaming GitHub - AuthBits/webmcp: A lightweight, prompt-driven MCP web research server for high-quality LLM powered information extraction. Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering Springdrift: An Auditable Persistent Runtime for LLM Agents with Case-Based Memory, Normative Safety, and Ambient Self-Perception High-Stakes Personalization: Rethinking LLM Customization for Individual Investor Decision-Making From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agents HUOZIIME: An On-Device LLM-enhanced Input Method for Deep Personalization TIDE: Token-Informed Depth Execution for Per-Token Early Exit in LLM Inference Characterizing WebGPU Dispatch Overhead for LLM Inference Across Four GPU Vendors, Three Backends, and Three Browsers LLM Targeted Underperformance Disproportionately Impacts Vulnerable Users
GitHub - kamaludu/bash4llm: Bash-first wrapper for Groq’s OpenAI-compatible API. Secure, portable, Termux-friendly.
kamaludu · 2026-06-28 · via Hacker News - Newest: "LLM"

Logo 320

CLI License: GPLv3 ShellCheck Smoke Tests

Bash4LLM⁺ — wrapper CLI sicuro, Bash‑first e completamente auditabile per l’API Chat Completions compatibile OpenAI di Groq (ed estendibile ad altri provider).

Bash4LLM⁺ è un singolo script Bash, auto‑contenuto, leggibile e verificabile.
Scaricalo, rendilo eseguibile, esporta la tua API key e inizia subito a usarlo.

Compatibile con ambienti Unix‑like: Linux, macOS, WSL, Cygwin, Termux (Android), BSD.


Caratteristiche principali

  • Lista modelli dinamica
    tramite GET https://api.groq.com/openai/v1/models
    → nessun modello hardcoded.

  • Sicurezza by design
    → nessun uso di /tmp, nessun eval, permessi restrittivi, validazione provider avanzata.

  • Struttura modulare a sezioni
    → PRECORE_BOOT, PRECORE_RUN, PROVIDER, CORE_SETUP, CORE_PROVIDER.

  • Sistema di Stato UI (ui_state)
    → il CORE espone costantemente metadati in formato JSON atomico per l'integrazione con GUI o strumenti esterni (es. Home Assistant).

  • Streaming e non‑streaming
    → output in tempo reale o completo a fine risposta.

  • Salvataggio automatico
    → per output lunghi oltre una soglia configurabile.

  • Gestione modelli avanzata
    → refresh, lista, default persistente, whitelist dinamica, auto‑selezione.

  • Extras opzionali
    → provider aggiuntivi (come Gemini, Hugging Face, Mistral), template, documentazione, strumenti di sicurezza.

  • Pronto per Termux / Android
    → rileva automaticamente l'ambiente Termux bypassando flock (spesso instabile o limitato a livello kernel/SELinux su Android) e devia trasparentemente la gestione della concorrenza sul robusto meccanismo di directory lock (mkdir atomico).


Modello di minaccia (versione breve)

Bash4LLM⁺ è progettato per ambienti single‑user (PC/laptop, server personali).

  • I provider sono codice eseguito nella tua shell: devono risiedere in directory sicure di tua proprietà.
  • Variabili come BASH4LLM_EXTRAS_DIR e BASH4LLM_TMPDIR sono considerate configurazione fidata.
  • Lo script non esegue mai l’output del modello.
  • I rischi TOCTOU e i limiti del parsing JSON/SSE sono mitigati e documentati.

Dettagli completi in SECURITY.


Requisiti

Bash4LLM⁺ richiede che i seguenti pacchetti (o equivalenti) siano disponibili nel PATH:

  • bash
  • coreutils
  • findutils
  • util-linux
  • gawk
  • curl
  • jq

Installazione

Tip

⏩ FAST FORWARD (Installazione Rapida)

Esegui questi comandi nel tuo terminale per avviare subito Bash4LLM⁺:

# 1. Clona il repository (solo l'ultimo commit per massima velocità)
git clone --depth 1 --branch main https://github.com/kamaludu/bash4llm.git repo-bash4llm  

# 2. Crea una cartella di lavoro ed estrai l'eseguibile
mkdir -p bash4llm
cp repo-bash4llm/bin/bash4llm bash4llm/
chmod +x bash4llm/bash4llm

# 3. Entra nella cartella e aggiorna i modelli 
cd bash4llm 
./bash4llm --refresh-models

Lo script ti chiederà l'inserimento della tua chiave API per il provider di default (Groq): Enter API key for provider groq (env GROQ_API_KEY):

Inserisci la tua API key, poi esportala per non doverla più inserire durante la sessione:

export GROQ_API_KEY="gsk_xxxxxxxxxxxxxxxxx"

Consigliato: installa gli Extras opzionali:

# 4. Installazione degli Extras
./bash4llm --install-extras ../repo-bash4llm/extras/

Usa Bash4llm ⚡

Istruzioni dettagliate in: INSTALL

In breve:

chmod +x bash4llm
export GROQ_API_KEY="gsk_xxxxxxxxxxxxxxxxx"
./bash4llm --help

Extras opzionali:

./bash4llm --install-extras

Con opzioni:

  • --source <dir>
  • --force
  • --dry-run
  • installazione selettiva:
    ./bash4llm --install-extras provider1 templateA

Uso rapido

Prompt diretto:

./bash4llm "scrivi una breve poesia in italiano"

Prompt multilinea:

./bash4llm <<'EOF'
scrivi una breve poesia
in italiano
EOF

Input da file:

Pipe:

echo "spiegami la relatività" | ./bash4llm

Modello specifico:

./bash4llm -m llama-3.3-70b-versatile "scrivi un saggio breve"

Dry run:

./bash4llm --dry-run "ciao"

Provider esterno (se installato):

./bash4llm --provider gemini "traduci questo"

Comandi, flag e opzioni disponibili

Modelli e provider

Flag Argomento Effetto
--refresh-models, --refresh-model no Aggiorna la lista modelli (richiede API key).
--list-models no Stampa lista modelli (formato interattivo).
--list-models-raw no Stampa lista modelli in formato raw (una riga per modello).
--list-providers no Stampa lista provider.
--list-providers-raw no Stampa provider in formato raw.
--set-default <model> Imposta modello di default persistente per il provider attivo.
-m <model>, --model <model> Imposta modello per questa esecuzione.
--provider <name> Imposta provider da CLI.
--provider no Se senza argomento → apre selezione interattiva.

Input (file, JSON, template, batch)

Flag Argomento Effetto
-f <file> Aggiunge file a FILE_INPUTS.
--json-input <json> Imposta input JSON (formato OpenAI-like).
--template <name> Applica template da BASH4LLM_TEMPLATES_DIR.
--batch <file> Esegue richieste batch (una riga = un prompt).

Sessioni

Flag Argomento Effetto
--session <id> Abilita sessione con ID specifico.
--session-window [n] opzionale Imposta finestra sessione (default 10 se non fornito).

Parametri modello / generazione

Flag Argomento Effetto
--system <text> Imposta system prompt.
--ture <n> Imposta parametro temperatura (da 0.0 a 2.0, alias canonico).
--temperature <n> Alias di --ture.
--max <n> Imposta max token.

Output e salvataggio

Flag Argomento Effetto
--save no Forza salvataggio output.
--nosave no Disabilita salvataggio.
--out <path> Percorso file/directory output.
--threshold <n> Soglia dimensione in byte per salvataggio automatico (default: 1000).
--json no Output JSON raw integro.
--pretty no Output JSON formattato.
--text no Output testuale standard estratto (comportamento predefinito).
--raw no Output testuale grezzo escludendo separazioni finali.

Modalità operative

Flag Argomento Effetto
--dry-run no Nessuna chiamata API reale (comportamento simulato).
--quiet no Riduce l'output non necessario e sopprime i titoli su TTY.
--stream no Streaming asincrono attivo.
--no-stream no Disattiva streaming asincrono.
--chat no Modalità chat interattiva REPL.
--bootstrap-only no Esegue solo validazione percorsi/lock e termina.

Configurazione e diagnostica

Flag Argomento Effetto
--show-config no Mostra configurazione completa attiva.
--diagnostics no Esegue diagnostica completa del sistema.
--version no Stampa versione dello script e termina.
-h, --help no Mostra help interattivo formattato da file.

Installazione extras

Flag Argomento Effetto
--install-extras opzionale Installa extras; può accettare directory sorgente.
--install-extras=<dir> Installa extras da directory sorgente specifica.

Terminazione parsing

Flag Effetto
-- Termina parsing opzioni.
-* Opzione sconosciuta → errore.
* Argomento posizionale → aggiunto a ARGS.

Configurazione e modelli

File di configurazione

  • $BASH4LLM_CONFIG_DIR/config
    → parametri locali (MODEL, TURE, MAX_TOKENS, FORMAT, THRESHOLD)

  • $BASH4LLM_CONFIG_DIR/model.$PROVIDER
    → modello predefinito per provider

  • $MODELS_FILE
    → whitelist modelli aggiornata da --refresh-models

Precedenza selezione modello

  1. -m/--model
  2. model.$PROVIDER
  3. auto‑selezione provider (auto_select_model_<provider>)
  4. prima voce della whitelist (models.txt)
  5. configurazione globale legacy config (MODEL=...)

File temporanei e output

  • Nessun uso di /tmp a livello di sistema operativo condiviso.
  • File temporanei isolati in directory $RUN_TMPDIR con permessi 700 (umask 077).
  • File salvati con permessi 600.
  • Con --out Bash4LLM⁺ crea la directory se possibile.

📁 Sistema di Stato UI (ui_state)

Bash4LLM⁺ espone metadati operativi destinati a GUI/strumenti esterni tramite file JSON atomici in:

$BASH4LLM_CONFIG_DIR/ui_state

Contiene:

  • sessions/<id>.json → stato sessione (active, msg_count, last_ts)
  • sessions/index.json → elenco sessioni
  • last_api.json → ultimo risultato API (http_status, req_id, edgecase_detected, ecc.)
  • last_history.json → ultimo salvataggio history
  • provider_capabilities.json → capacità provider attivo (streaming, refresh_models)

La GUI (extra opzionale) legge solo questi file per i placeholder CGI.


📘 Memoria contestuale in Bash4LLM⁺

Bash4LLM⁺ non mantiene memoria da solo.
La memoria esiste solo se attivi una sessione tramite --session.

Ogni sessione crea un file NDJSON persistente:

$BASH4LLM_HISTORY_DIR/sessions/<session_id>.ndjson

E Bash4LLM⁺ mantiene i metadati della sessione in:

$BASH4LLM_CONFIG_DIR/ui_state/sessions/<session_id>.json

Questi metadati sono la fonte canonica per GUI/strumenti esterni.


🟩 Uso corretto di --session

./bash4llm --session chat1 "Ciao"
./bash4llm --session chat1 "Riassumi ciò che ho detto"

🟩 Uso corretto di --session-window

./bash4llm --session chat1 --session-window 10 "continua"

🟧 Regola fondamentale

Per avere memoria contestuale devi sempre includere --session <id>.


Note di sicurezza

  • Nessun eval.
  • Nessuna esecuzione dell’output del modello.
  • Provider = codice: mantieni extras/providers sicuro.
  • Variabili d’ambiente = configurazione fidata.
  • TOCTOU mitigato.

Codici di uscita

Codice Variabile Significato
0 - Successo
10 BASH4LLM_ERR_NO_API_KEY API key mancante
11 BASH4LLM_ERR_BAD_MODEL Modello non valido o non in whitelist
12 BASH4LLM_ERR_CURL_FAILED Errore rete/curl
14 BASH4LLM_ERR_NO_PROMPT Nessun prompt fornito
15 BASH4LLM_ERR_TMP Errore generico filesystem / temporanei
16 BASH4LLM_ERR_API Errore HTTP/API del fornitore

Variabili principali

Variabile Necessaria Descrizione
GROQ_API_KEY sì per chiamate API API key provider Groq.
BASH4LLM_CONFIG_DIR consigliata Directory configurazione.
BASH4LLM_MODELS_DIR consigliata Directory modelli.
BASH4LLM_TMPDIR Directory temporanea.
BASH4LLM_HISTORY_DIR consigliata Directory sessioni e cronologia.
MODEL no Modello attivo.
PROVIDER no Provider attivo.
ALLOWED_MODELS no Whitelist modelli ammessi.

Licenza

Bash4LLM⁺ è distribuito sotto licenza GPL v3.
Vedi LICENSE.


Contatti

Autore: Cristian Evangelisti
Email: opensource​@​cevangel.​anonaddy.​me
Repository: https://github.com/kamaludu/bash4llm