Developer Builds Claude Code Hook to Auto-Log and Analyze AI Agent Usage

·1 views

A developer has created a logging system for Claude Code that automatically records every AI agent invocation using the platform's Stop hook and transcript_path payload. When a session ends, a chain of shell scripts triggers a Python parser that reads the conversation transcript and appends structured records to a JSONL log file. The system has so far captured 553 agent invocations, revealing that 47 defined agents had never been called even once. Unused agent definitions are considered actively harmful because they add unnecessary context to every session. The tool uses a combination of session ID and tool-use ID as a uniqueness key to prevent duplicate entries across multiple Stop hook firings within the same session.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Dev Tutorial Explains How to Sync YouTube Playlists to Spotify Using Python

A developer tutorial published on DEV Community walks readers through writing Python code to automatically sync a YouTube playlist to Spotify. The project uses three Python files — spotify.py, youtube.py, and script.py — alongside a .env file that stores private API credentials locally. The YouTube module handles paginated API responses, fetching up to 50 videos per request to retrieve full playlists exceeding that limit. A parsing function splits video titles on hyphens to identify possible artist and song combinations, since YouTube creators follow no standard naming format. A separate cleaning function strips common YouTube title clutter such as 'Official Music Video' and '[HD]' before sending search queries to Spotify.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Browser Tool Tests LLM Security Using Deterministic Grading Without an AI Judge

A developer has released The AI Crash Test, a browser-based tool that evaluates large language models against adversarial prompts using deterministic, code-based grading rather than a secondary AI judge. Users supply their own API key, which is sent directly to the model provider and never passes through the tool's server — a claim the developer says anyone can verify via browser DevTools in about 30 seconds. The grading engine, called gradecore, produces byte-identical scores across repeated runs, eliminating variability from temperature or judge drift. The tool covers eight tasks across seven attack categories, including prompt injection, hallucination baiting, and refusal calibration, generating a severity-weighted vulnerability report. The developer acknowledges that more comprehensive red-teaming tools such as NVIDIA's garak and Microsoft's PyRIT exist, positioning this project not as a new category but as a narrower, auditable alternative focused on reproducibility.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Manticore Search 27.1.5 Adds Native Authentication and Authorization Controls

Manticore Search has introduced built-in authentication and authorization starting with the 27.x release line, officially available in version 27.1.5. The feature supports SQL and HTTP clients, with login via username/password, Basic auth, or Bearer tokens. A five-action permission model — read, write, schema, replication, and admin — can be scoped to specific tables or wildcards, letting operators grant each user only the access they need. Previously, teams relied on external tools like nginx to manage access control, but the new system brings user and permission management natively into Manticore via SQL commands. The update is aimed at teams running shared or customer-facing search infrastructure where different services, tenants, or roles require distinct trust levels.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

React Native Explained: One Codebase for iOS and Android Apps

React Native, open-sourced by Facebook in 2015, allows developers to build a single codebase that runs natively on both iOS and Android platforms, cutting development costs roughly in half compared to building two separate apps. Unlike older web-based cross-platform tools, React Native compiles to actual native UI components, meaning apps feel and perform like true native applications. Major companies including Instagram, Shopify, Microsoft, and Coinbase have relied on the framework in production for nearly a decade. Experts stress that thorough requirements planning before development begins is critical, as skipped or rushed planning is cited as the most common reason mobile app projects fail. Developers also note that iOS and Android users have distinct behavioral expectations around navigation and interface conventions, making platform-specific design adaptation important even when sharing a single codebase.

0 comments Read more at DEV Community