Releases·Tool Release
Show HN: Local audit trail for Claude Code tool calls with rule-based risk analyzer
A developer shared a GitHub project providing a local audit trail for Claude Code tool calls, paired with a rule-based risk analyzer. Users can supply a llama-cpp-compatible GGUF model to get a local second opinion on the risk of tool calls, keeping the whole workflow on-device.
3 engagement2 sourcesSun, Sep 6, 2026, 06:43 PM
AI Overview
- Logs Claude Code tool calls to a local audit trail and runs a rule-based risk analyzer for a second opinion
- Accepts any llama-cpp compatible GGUF model to assess tool-call risk locally
Entities
Related
Releases·Tool Release
Developer launches Conduct, open-source guardrails for LLM and MCP tool calls
A developer shared Conduct on Hacker News, an open-source tool providing guardrails for LLM and MCP tool calls. It gives practitioners a way to validate and control tool invocations in agent pipelines. The project is hosted on GitHub.
A developer shared Conduct on Hacker News, an open-source tool providing guardrails for LLM and MCP tool calls. It gives practitioners a way to validate and control tool invocations in agent pipelines. The project is hosted on GitHub.
26 engagement1 sourcehackernewsFri, Aug 28, 2026, 07:29 PM
Products·launched
Sentrint: Security scanner for LLM-built projects
Sentrint is a security scanner designed for projects built with LLMs. It helps developers identify vulnerabilities specific to AI-generated code, addressing the unique security challenges introduced by LLM-assisted development.
Sentrint is a security scanner designed for projects built with LLMs. It helps developers identify vulnerabilities specific to AI-generated code, addressing the unique security challenges introduced by LLM-assisted development.
7 engagement1 sourcehackernewsThu, Aug 20, 2026, 12:46 PM
Releases·Tool Release
Show HN: Tracelint – a linter for AI agent traces, no LLM judge
Tracelint is a new open-source linter that statically analyzes AI agent traces to catch errors without relying on an LLM judge. It provides deterministic checks for agent workflows, aiming to improve reliability and debuggability.
Tracelint is a new open-source linter that statically analyzes AI agent traces to catch errors without relying on an LLM judge. It provides deterministic checks for agent workflows, aiming to improve reliability and debuggability.
1 engagement1 sourcehackernewsTue, Aug 18, 2026, 02:49 PM
Products·launched
Observability for coding agents and LLM applications
Telemetry.dev provides observability tools for coding agents and LLM applications, helping developers monitor, debug, and optimize AI-powered workflows. It solves the problem of understanding agent behavior and performance in production.
Telemetry.dev provides observability tools for coding agents and LLM applications, helping developers monitor, debug, and optimize AI-powered workflows. It solves the problem of understanding agent behavior and performance in production.
2 engagement1 sourcehackernewsTue, Jul 21, 2026, 07:33 PM
Social·Community
Claude built my 3-Mac-mini local LLM cluster. I set the rules and checked results. Here's what broke and what I lost.
A homelab builder used Claude Code as a supervised background agent to design, benchmark, and configure a three-Mac-mini cluster serving one 35B MoE model across two machines. The write-up documents working delegation patterns, concrete benchmarks, and the failures Claude hit. It matters to practitioners as a case study in agent-led infrastructure setup with human rule-setting.
A homelab builder used Claude Code as a supervised background agent to design, benchmark, and configure a three-Mac-mini cluster serving one 35B MoE model across two machines. The write-up documents working delegation patterns, concrete benchmarks, and the failures Claude hit. It matters to practitioners as a case study in agent-led infrastructure setup with human rule-setting.
5 engagement1 sourceredditThu, Oct 1, 2026, 07:14 PM