Search Results for "injection"

Codex-5.5-codex-instruct-5.5 is a high-risk prompt-injection utility aimed at modifying how Codex CLI behaves. The repository presents itself as an early command-line version of a jailbreak-style tool for injecting unrestricted-mode instructions. It claims to use an official configuration mechanism rather than binary modification, network interception, or process tampering. The project includes a Python script, an examples folder, assets, a README, and an MIT license.

...It lets teams define controls once and apply them consistently to agents without rewriting the agent’s core code. The platform evaluates agent inputs and outputs against configurable policies to reduce risks such as prompt injection, unsafe responses, sensitive data exposure, and policy drift. It is designed for production environments where organizations need observability, enforcement, and governance around autonomous or semi-autonomous AI systems. The repository includes SDKs, a server, telemetry components, examples, and integrations for common agent frameworks. ...

...Visually track events such as LLM calls, tools, and multi-agent interactions. Rewind and replay agent runs with point-in-time precision. Keep a full data trail of logs, errors, and prompt injection attacks from prototype to production. Native integrations with the top agent frameworks.

Agent SOP is a framework that implements structured operational procedures (SOPs) for autonomous agents so that they can carry out complex multi-step tasks reliably and in a defined order. Instead of relying solely on broad language model reasoning, this project enforces explicit step sequences with checkpoints, conditional transitions, and rollback logic, making agent workflows more predictable and auditable. It defines reusable SOP templates that agents can instantiate with...

iFixAi is an independent auditing toolkit for evaluating whether an AI agent performs the business job it was assigned. Instead of measuring only latency, token use, or prompt-injection resistance, it examines operational behavior, organizational alignment, and task outcomes. A standard run performs 32 inspections across five evaluation pillars and produces an A-to-F grade. Users can configure and launch audits through a guided CLI, explicit command flags, or an agent plugin and skill. The same engine can test hosted providers or a real agent endpoint and can use self-grading, an independent judge, or a multi-judge ensemble. ...

...The assistant offers features such as voice control using fully local speech-to-text (Whisper) and text-to-speech (Piper) capabilities, news aggregation with extractive summarization, and smart home or Bluetooth device control. SafeClaw supports multiple channels, including CLI and Telegram, and avoids prompt injection risk because it doesn’t rely on LLMs for core operations.

pg-aiguide is a tool designed to enhance AI-assisted development with PostgreSQL by providing structured knowledge and skills directly to coding agents. It acts as a bridge between database documentation and AI tools, enabling more accurate generation of SQL queries and database interactions. The system integrates with Claude Code through an MCP server, allowing agents to access curated PostgreSQL knowledge in real time. It focuses on improving developer productivity by reducing errors and...

HunterX AI-Assisted Offensive Security Engine Discover. Validate. Prove. Report. HunterX is not merely a vulnerability scanner. It is an open-source, AI-assisted offensive security engine for planning and running authorized security-assessment missions. It combines reconnaissance, security-tool orchestration, AI-assisted reasoning, hypothesis-driven investigation, vulnerability validation, evidence collection, proof / PoC engineering, replay / reproducibility, correlation, impact...