Snapshot
Continue to About ↓- What it does
- iFixAi is an independent auditing service for deployed AI agents.
More detail
- It connects to an agent through GitHub or MCP, builds a simulation environment, sends test and adversarial inputs to the agent's endpoint, and returns an audit report showing where the agent's behaviour diverges from its assigned purpose, permissions, workflows, roles and evidence trail. iFixAi positions the report as diagnostic evidence, not certification or a security guarantee.
- Best for
- Enterprises and startups running deployed AI agents that have real authority over money, data or customers, especially teams already working in GitHub or MCP tooling who want an outside view of whether an agent stays inside its permissions, approvals and role boundaries.
- Pricing
- Hosted auditing is sold per agent, per month in tiers: Startup (2 audits/reports per month, up to 3 agents), Growth (4 audits/reports per month, up to 3 agents, 'Audited by iFixAi' included) and Enterprise (8 audits/reports per month, up to 3 agents, 'Audited by iFixAi' included).
More detail
- Agentic Enterprise
- Main limitation
- Buyers who need a certification, compliance sign-off or security guarantee, or teams that want a fully managed service without exposing agent code or prompts to an external auditor.
More detail
- Also a poor fit for anyone needing a published per-tier price before committing, since the hosted tiers do not list dollar amounts publicly.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- iFixAi is an independent auditing service for deployed AI agents.
- It connects to an agent through GitHub or MCP, builds a simulation environment, sends test and adversarial inputs to the agent's endpoint, and returns an audit report showing where the agent's behaviour diverges from its assigned purpose, permissions, workflows, roles and evidence trail. iFixAi positions the report as diagnostic evidence, not certification or a security guarantee.
What it helps with
- Standard evals, observability and benchmarks measure whether an agent completed its task, not whether it stayed inside i
- An agent can pass every eval and still hide information, override a declined decision, act without authorisation or expo
- Teams lack independent, reproducible evidence of an agent's behaviour to defend to stakeholders or auditors.
- Agents that change continuously need repeated proof rather than a one-off check.
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| GitHub connection | Yes | Reads the agent's code to build the simulation environment; iFixAi says it only reads what you connect and that code and prompts stay with you. |
| MCP connection | Yes | One prompt pasted into Claude Code, Codex, Cursor or VS Code connects and starts the audit. |
| AGENTS.md auto-detection | Yes | Repositories containing an AGENTS.md are picked up and ready to connect. |
| Simulation environment review before audit | Yes | The environment built from the agent definition can be reviewed before any audit runs. |
| Adversarial and test input probing | Yes | Sends test inputs to the agent's endpoint and records replies and actions. |
| Findings report with evidence | Yes | AI models review results and produce findings with supporting evidence. |
| A–F grade across five core pillars | Yes | Weighted average of Fabrication, Manipulation, Deception, Unpredictability and Opacity; mandatory minimum checks cap the score at 60% if missed. |
| Framework mapping | Limited | Findings can be mapped to EU AI Act, NIST AI RMF, ISO/IEC 42001 and OWASP Top 10 for LLM Applications, but compliance remains the customer's determination. |
| Certification or security guarantee | No | iFixAi presents audits as diagnostic evidence, not certification; a badge means an audit occurred, not that the agent passed. |
| Published hosted subscription prices | No | Tier structure is public but dollar amounts per agent per month are not. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- ifixai.ai
Technical profile details
Record ID
36ef8bd2-543d-46c4-b5ad-a5fda21b2f72Canonical URLhttps://profiles.scribble.network/listing/ifixai-36ef8bd2