Shivom Sharma
ReportIndependent security researcher based in India and the sole creator and maintainer of tirith. My professional background is adversarial security research and risk analysis: Research & Advisory Lead at Impossible Finance, Research Lead at Lemniscap, and DeFi R&D at Nethermind - roles centered on studying attacker behavior, reviewing code and system designs for exploitable assumptions, and adversarial threat modeling against well-resourced attackers. I designed and built tirith from scratch in Rust: the three-tier analysis pipeline, the shell hooks, command and paste analysis, URL and package extraction, the policy system, the Ed25519-signed threat database with rollback protection, credential redaction and DLP, AI-agent config scanning, the MCP server and gateway, and all 110+ detection rules across 16 categories. Because I wrote the system end to end, I know its failure modes in detail - which is what makes hardening, false-positive reduction, and new detection fast and correct rather than guesswork. I build local-first, telemetry-free security tooling for the people who cannot deploy heavy commercial endpoint security: individual developers, AI-agent users, and high-risk open-source maintainers. This focus shows in practice: tirith ships a frequent, cosign-signed release cadence, daily Ed25519-signed threat-database builds, a fuzzing harness, a public threat model with explicit non-goals, and an internal security review with regression tests already merged. I am bringing an already-adopted, shipping tool rather than a concept on paper.
tirith
tirith is a local, open-source, pre-execution security layer written in Rust. It inspects a command, paste, package install, or AI-agent/MCP tool call BEFORE it executes - the exact moment modern supply-chain and social-engineering attacks land (malicious and typosquatted packages, pipe-to-shell installers, poisoned CI, and prompt-injected AI coding agents). It flags homograph/IDN-spoofed URLs, terminal-injection (ANSI/bidi/zero-width) and steganographic deception, pipe-to-shell and base64 decode-execute chains, credential exfiltration, malicious packages and infrastructure, and unsafe AI-agent or MCP configurations. It installs as a shell hook (bash, zsh, fish, PowerShell; nushell warn-only) and as an inspection engine callable from the CLI, CI, and AI agents. A three-tier pipeline (sub-millisecond fast gate, then a tokenizer, then 110+ detection rules across 16 categories) is backed by a daily Ed25519-signed local threat database built from OpenSSF Malicious Packages, Datadog Security Labs, abuse.ch, ecosyste.ms, and CISA KEV. It also ships an MCP gateway that vets AI-agent shell tool calls and a verified download-and-execute alternative to 'curl | bash'. It runs offline by default and sends no telemetry. No other widely used open-source tool combines pre-execution inspection across shells, pastes, package installs (nine ecosystems), and AI-agent/MCP configs in a single signed, offline, AGPL binary. Traction is organic and unsolicited: 2,400+ GitHub stars, 80+ releases, distribution across nine packaging channels with cosign-signed releases, and integration into Nous Research's Hermes Agent (193,000+ stars) - merged by that project's own lead, not by me - as its pre-execution command-vetting layer. Built and maintained in India, tirith stays free for everyone; grant support funds expanded AI-agent and MCP detection, hardening, and a public benchmark of unsafe agent commands.
tirith one-year development grant
Funds a year of security engineering on tirith: expanded AI-agent/MCP detection, a hardened MCP gateway, an agent+shell compatibility matrix behind a CI gate, signed-threat-DB hardening, parser fuzzing, a public unsafe-agent-command benchmark, Windows support, and docs. Local-first and telemetry-free, so the full amount funds engineering time - an efficient ask for an already-adopted, shipping tool. Sustained via dual-licensing.
Monthly sponsorship
Recurring monthly support for tirith's continued maintenance, daily signed threat-database updates, and security response.
One-time contribution
Any one-time contribution toward tirith's development, maintenance, and security work.
| Year | Income | Expenses | Taxes | Description |
|---|---|---|---|---|
| 2026 | 0 INR | 0 INR | 0 INR | Bootstrapped and maintained entirely on unpaid maintainer time since launch in February 2026; no external grant or sponsorship funding received to date. Running costs are minimal by design - tirith is local-first with no servers or telemetry, so the only recurring spend is CI, release signing, and threat-feed refresh built on free public sources. |