Skip to content
View Iceman2050's full-sized avatar
0 followers · 2 following

Block or report Iceman2050

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Iceman2050/README.md

Ingenious AI — AI security and agent-safety tooling

Aldo Guerra | Founder of Ingenious AI

I build practical security tools for AI agents, MCP integrations, developer infrastructure and software supply chains.

I'm the founder and technical builder behind Ingenious AI.

My work focuses on turning emerging security problems into small, useful, reproducible tools that developers can actually run — zero-dependency where possible, honest about what a check can and cannot prove.

What I'm working on

  • AI Agent Security — Authorization, least privilege, action controls and evidence.
  • MCP Security — Server provenance, tool permissions, configuration auditing and trust.
  • Software Supply Chain Security — Package integrity, plugin verification and CI/CD protection.
  • AI Security Testing — Prompt injection, context boundaries, RAG poisoning and agent behavior.
  • Security Automation — Lightweight tools, CLI utilities and reusable GitHub Actions.

Start here

  • mcp-permission-firewall — Audit and harden MCP server configs: least-privilege policies for AI agent tools, before you trust a pasted config.
  • cve-to-checker — Turn vulnerability advisories into reusable "am I affected?" exposure-checker projects.
  • cve-2026-110447-checker — A focused, passive Nextcloud exposure checker for CVE-2026-110447.
  • ghostaction-guard-action — A free GitHub Action that scans your workflows for GhostAction-style malicious CI injection.
  • plugin-provenance-verifier — Audit where your Claude Code plugins actually came from — and catch silent drift.
  • history-tamper-guard — A user-side notary for agent conversation history: hash-chained turns, tamper detection.
  • api-key-sprawl-mapper — Maps API-key distribution across repositories without ever printing secret material.
  • saas-offboarding-checklist — Ordered SaaS access-revocation runbooks for departing employees.

The full open-source estate — every tool, one registry-verified directory: Ingenious AI — All Links

The hosted platforms

The free tools solve one problem at a time. The Ingenious AI platforms solve the recurring, operational version of the same problems — continuous control, evidence and monitoring:

  • Agent Auth Gateway — spend caps, allow/deny policies and approval queues for AI agents.
  • Agent Flight Recorder — hash-chained, tamper-evident records of agent actions.
  • BreachHunter — automated website security scanning; the free front door.
  • DomainGuard — lookalike-domain and phishing-impersonation defense.

Engineering principles

  • Prefer simple, understandable implementations.
  • Minimize unnecessary dependencies.
  • Make security limitations explicit.
  • Provide reproducible examples and test results.
  • Respect privacy and minimize data collection.
  • Document authorized-use requirements.
  • Never claim a security check proves more than it actually verifies.

Get involved

Issues, technical feedback and constructive contributions are welcome.

Building useful security infrastructure, one practical tool at a time.

Pinned Loading

  1. mcp-permission-firewall mcp-permission-firewall Public

    Audit and harden MCP server configs — least-privilege policies for AI agent tool access. By Ingenious AI.

    Python

  2. cve-to-checker cve-to-checker Public

    Open framework: turn any CVE advisory into an 'am I affected?' checker page in minutes. By Ingenious AI.

    Python

  3. cve-2026-110447-checker cve-2026-110447-checker Public

    Free CLI: check a Nextcloud instance for CVE-2026-110447 exposure. Stdlib only, honest verdicts. By Ingenious AI.

    Python

  4. ghostaction-guard-action ghostaction-guard-action Public

    Free GitHub Action: scan your workflows for malicious CI injection (GhostAction campaign). 8 detection rules, zero dependencies. By Ingenious AI.

    Python

  5. plugin-provenance-verifier plugin-provenance-verifier Public

    Audit where your Claude Code plugins actually came from — and catch silent drift.

    Python

  6. history-tamper-guard history-tamper-guard Public

    Conversation-history notary for agent CLIs (2026-09-24 Darktrace hook): hash-chains turns as written, detects retroactive edits & fabricated turns

    Python