Security

Practical reporting on AI systems, tools, and engineering execution.

Infisical: an open-source platform that bundles secrets, PKI, and agent-based secret delivery for teams

Infisical is an open-source secrets and certificate platform that centralizes secret versioning/rotation, dynamic secrets, Kubernetes delivery, and private PKI while offering both self-hosted and cloud deployment options.

Git repo

October 12, 2026 2m

PyRIT: Microsoft’s open-source Python framework for identifying generative‑AI risks

PyRIT is an open-source Python framework from Microsoft intended to help security teams and engineers identify risks in generative‑AI systems; review the repo website and whitepaper to assess fit and limitations before adoption.

Git repo

October 12, 2026 2m

HealthClaw Guardrails — open-source FHIR × MCP guardrail layer (v2.0.0)

An MIT-licensed reference implementation that places a guardrail layer between AI agents and FHIR servers, with connectors, adapters, and test coverage intended for experimentation and integration work.

Git repo

October 12, 2026 1m

CoWork OS: a local-first desktop that runs agents, CLIs, browsers and multiple model providers in one governed workspace

CoWork OS is an open-source desktop workspace that unifies coding, browser testing, documents, automations and long‑running agents while letting you route work to local models or cloud providers under a governed, inspectable runtime.

Git repo

October 12, 2026 3m

sofagent — an open-source FDE harness to audit and constrain AI coding agents

An MIT-licensed harness layer that injects audit, snapshot, and rollback controls around AI coding agents; includes a CLI audit command, a sandbox demo, and documentation for integration.

Git repo

October 12, 2026 2m

Drix10/ai-resources — continuously generated, categorized AI systems archive (maintainer: Drix10)

A continuously regenerated, categorized archive of AI systems engineering resources and authored breakdowns, driven by an autonomous pipeline that publishes distilled articles to this repo and syndicated blogs.

Git repo

October 12, 2026 2m

OpenOSINT — an AI-driven OSINT agent with REPL, MCP server, CLI, Web UI, and Apify cloud actors

OpenOSINT exposes 21 OSINT tools behind a natural-language AI interface plus an MCP server; use it locally (REPL/CLI/Web) or as pay-per-use Apify actors for quick runs.

Git repo

October 12, 2026 2m

Strix — open-source, Docker-based AI pentesting agents that produce exploit PoCs

Strix is an open-source tool that runs local, Docker-isolated AI penetration tests using your LLM key, producing working proof‑of‑concepts and suggested fixes for CI/CD integration and developer workflows.

Git repo

October 12, 2026 2m

NeverC: a single-binary C23 toolchain with built-in runtimes, DynCode pipeline, and plugin API

NeverC packages a C23 compiler, linker, runtimes and a 130‑phase plugin API into one LLVM-based binary with cross-compilation and a DynCode pipeline — useful for security research, with caveats for production use.

Git repo

October 12, 2026 3m

0day‑Rubbish — AI-driven mass disclosure of verified 0‑day PoCs (repo teardown)

A repository that claims to auto-discover and publish verified, working 0‑day proof‑of‑concepts weekly using LLM ensembles — potentially useful for threat modeling but requiring legal, ethical, and operational vetting.