# AI Agents · Tech watch [Canonical source]() Systems using models to choose actions, call tools and perform tasks. ## [PGX: disposable PostgreSQL databases for agents and tests]() PGX runs isolated PostgreSQL databases inside a single Rust runtime, targeting short-lived agent, CI, and test environments. On-demand creation from a sealed template and copy-on-write on ZFS reduce the cost of branching; without block cloning, the system makes a full copy. The project is at **v0.1.0** and targets disposable databases, not production data: `fsync` is disabled, and one runtime is a single failure domain. Its benchmarks document their measurements and conditions; the PostgreSQL regression suite reports zero result differences in the stated configuration, without claiming general compatibility. 2026-10-22 · [Markdown]() · [Original resource]() ## [REA: agent-powered reverse engineering for binaries and applications]() REA offers a unified approach to reverse engineering, driven by agents that can analyze native binaries, JavaScript/Electron apps, websites, or databases without source code access. The tool integrates with agents like Claude Code or Cursor, provides evidence and limitations for each analysis, and leverages native tools (Hopper, Ghidra, IDA) for deep inspection. The documentation details supported scenarios, provider setup, and CLI workflows, with concrete examples of feature-level reverse engineering from real applications. 2026-10-19 · [Markdown]() · [Original resource]() ## [Engram: persistent memory for MCP-compatible AI agents]() Engram provides **persistent memory** for AI agents, compatible with any agent supporting the MCP protocol (Claude Code, Gemini CLI, Codex, VS Code, etc.). The tool runs as a single Go binary with no external dependencies, storing memories in a SQLite file with built-in FTS5 search. It features a TUI, Git sync, and a self-hostable cloud extension, while remaining local-first and open source under the MIT license. 2026-10-16 · [Markdown]() · [Original resource]() ## [Medula: inspect agent interactions inside an app]() Medula provides development tools for AI agents and documents how to connect them to Vue, React, Svelte, and Solid apps using Vite 8. The repository also includes examples and end-to-end tests for observing how an agent interacts with an application at runtime. 2026-10-09 · [Markdown]() · [Original resource]() ## [An evidence-graded index of the Jev ecosystem]() Awesome Jev Live indexes SDKs, MCP tools, agents, apps, and open models related to TypeSafe System One. Entries are graded by evidence and organized by category and language; the catalog rebuilds every two hours to track changes across the ecosystem. 2026-10-05 · [Markdown]() · [Original resource]() ## [uniTerm: a cross-platform terminal with an AI agent]() uniTerm brings more than 30 protocols into a cross-platform terminal, including SSH, RDP, SFTP, databases, and Kubernetes. The repository also describes an AI agent that can plan and run multi-turn commands, plus workflows combining remote access, terminal sessions, and file transfer. 2026-10-05 · [Markdown]() · [Original resource]() ## [A library of specialized agents for software teams]() Agency Agents collects specialized agents with defined personalities, processes, and deliverables across design, engineering, product, security, and marketing. The repository explains several ways to use them, including Claude Code and other AI-assisted development tools. 2026-09-06 · [Markdown]() · [Original resource]() ## [Audit a website’s visibility in AI-powered search]() This Claude Code skill audits how a site appears in AI-powered search. It combines citability scoring, AI crawler analysis, brand-mention checks, and structured-data reviews, then produces a report teams can use to prioritize improvements. 2026-09-01 · [Markdown]() · [Original resource]() ## [A compact web-browsing CLI for AI agents]() A compact CLI turns websites into commands tailored for AI agents, returning the useful information without loading tens of thousands of tokens. The repository documents its commands and supported sites, making it a practical way to reduce the cost of automated browsing. 2026-08-24 · [Markdown]() · [Original resource]() ## [Build a knowledge graph for coding assistants with Graaph]() Graaph builds queryable graphs from code, docs, papers, and diagrams to give coding assistants more context. Its site demonstrates sample graphs and navigation tools. 2026-08-03 · [Markdown]() · [Original resource]() ## [Start a spec-driven development workflow with GitHub Spec Kit]() Spec Kit provides tools for turning requirements into specifications, plans, and implementation tasks. Its repository documents a workflow that puts written decisions before code generation. 2026-07-15 · [Markdown]() · [Original resource]() ## [Strix: automate AI-assisted penetration testing]() Strix is an open-source tool that uses agents to test an application and report vulnerabilities. Its repository exposes the workflow components for evaluation in an authorized environment. 2026-06-06 · [Markdown]() · [Original resource]() ## [Emil Kowalski’s skill for polished interface details]() This skill turns Emil Kowalski’s motion and interface-craft guidance into instructions an agent can use. It focuses on transitions, states, and details that make interactions clearer. 2026-05-22 · [Markdown]() · [Original resource]() ## [Help agents produce less generic interfaces with Taste Skill]() Taste Skill collects open-source guidance for an agent’s visual decisions. Installing it adds a design framework beyond default components and styles. 2026-05-22 · [Markdown]() · [Original resource]() ## [Run a project pre-mortem before implementation]() This skill guides a pre-mortem session that imagines a project failure to surface risks early. The Markdown file can be used by a coding agent or in product discovery. 2026-05-18 · [Markdown]() · [Original resource]() ## [Structure an agent’s software workflow with Superpowers]() Superpowers provides skills for breaking down work, exploring a problem, and checking an implementation. Its repository describes a reusable method for multiple coding agents. 2026-05-18 · [Markdown]() · [Original resource]() ## [Give coding agents a design vocabulary with Impeccable]() Impeccable provides design commands and guidance to help an agent refine an existing interface. It covers typography, color, and consistency with a component system. 2026-05-10 · [Markdown]() · [Original resource]() ## [Prototype interfaces with your coding agent using OpenDesign]() OpenDesign is an open-source workspace for creating prototypes, pages, and presentations with a coding agent. It centers the workflow on developer files and tools. 2026-05-10 · [Markdown]() · [Original resource]() ## [Map a repository to guide code reviews with Code Review Graph]() Code Review Graph builds a local graph of a codebase and exposes it through MCP and CLI tools. It aims to give agents relevant context without loading the entire repository every time. 2026-03-23 · [Markdown]() · [Original resource]() ## [Bring current library docs into coding agents with Context7]() Context7 gives coding agents documentation snippets for a specific library and version. It helps reduce answers based on outdated or guessed APIs. 2026-02-26 · [Markdown]() · [Original resource]() Page 2 / 3 [Previous]() [Next]()