iFixAi
Independent Auditing of AI Agents. Run by human or the agent itself, to answer the most crucial question in the AI Agent Economy. Is the agent doing what is supposed to do? With iFixAi you can have this answer in less than 120 seconds.
iFixAi is an independent AI agent auditing skill that evaluates whether agents perform their intended business functions correctly. It runs 49 inspections across five core pillars—fabrication, manipulation, deception, unpredictability, and opacity—delivering an A-F grade in under 120 seconds. The skill can be installed as a native slash command in Claude Code, Codex, Cursor, VS Code, Windsurf, and other coding assistants, or run via CLI with guided setup or explicit flags. It supports testing both bare model APIs and real deployed agents through HTTP endpoints or custom adapters.
Key Features
Use Cases
- 01Audit AI coding assistants and development agents before production deployment to verify business compliance
- 02Run automated agent governance checks in CI pipelines using explicit CLI flags
- 03Validate that customer service agents follow organizational policies and don't fabricate claims
- 04Detect privilege escalation, prompt injection, and policy drift in production agent deployments
- 05Compare agent performance across different LLM providers using independent cross-vendor judges
- 06Verify agent audit trails, human escalation paths, and regulatory compliance requirements
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
iFixAi — FAQ
What is iFixAi?+
iFixAi is an independent auditing tool for AI agents that tests whether they perform their intended business functions correctly. It runs 49 inspections across five core pillars and delivers an A-F grade, distinguishing itself from traditional evals by focusing on operational assurance and business KPIs rather than just technical metrics like latency or token efficiency.
How do I install iFixAi as an agent skill?+
Run 'uvx ifixai install --agents <agent-name>' to scaffold a /ifixai-skill slash command into any supported coding assistant (claude, codex, cursor, vscode, windsurf, cline, continue, gemini, or zed). For Claude Code or Codex, you can also install as a native plugin using '/plugin marketplace add ifixai-ai/iFixAi' followed by '/plugin install ifixai@ifixai-community'.
Which AI clients does iFixAi work with?+
iFixAi works with Claude Code, Codex, Cursor, VS Code with Copilot, Windsurf, Cline, Continue, Gemini, and Zed as a skill or plugin. It can also run as a standalone CLI tool or test any agent accessible via HTTP endpoint.
Do I need API keys to use iFixAi?+
Yes, you need API keys for both the agent being tested (system under test) and the judge model that grades it. For citable results, these should be from different providers. The guided wizard auto-detects keys in your environment, or you can pass them via flags or environment variables.
Is iFixAi free to use?+
Yes, iFixAi is open source under Apache 2.0 license and free to use. However, you'll incur API costs from the LLM providers for the models being tested and the judge models. A full suite run with recommended judges costs approximately $10-18 depending on the judge configuration.
What prerequisites does iFixAi require?+
iFixAi requires Python 3.10 or higher. For the CLI, install via 'pip install ifixai[provider]' where provider matches your target (openai, anthropic, gemini, etc.). For the skill installation method, you only need uv and Python 3.10+.
How do I install iFixAi?+
Open the source repository on GitHub and follow its README. iFixAi is a skill — MCP Agents Market links you directly to the official repo.
Is iFixAi free?+
iFixAi is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.