Overview
Tessl is a governance platform for coding agents. It turns your team's own coding standards into versioned, reusable skills and enforces them on every code change. It sits between a linter and AI code review: a linter only checks deterministic rules, AI review sweeps a whole PR for general problems, and Tessl's Verifiers turn the team rules that neither can reliably check into a pass-or-fail gate that runs like lint or tests.
On October 15, 2026, Tessl released Verifiers into general availability. Rules live as JSON files in your repo, versioned alongside your code, and an LLM judge scores each one pass or fail on every change. Beyond Verifiers, the platform ships a skill Registry and a set of automated Loops covering code review, skill optimization, security scanning, doc maintenance, model evaluation and cost benchmarking, all aimed at moving coding agents from writing code to shipping it against your standards.
Key Features
- Verifiers gate: JSON rules in your repo scored pass or fail by an LLM judge on every change, runnable locally, in pre-commit, or as a CI gate
- Fills the linter gap: Handles team standards a linter cannot express, like user-facing errors must tell the user what to do next
- Team-owned Code Review: Reviews the whole PR against your own lens skills and remembers what is fixed, declined, or still open
- Skill Registry: A searchable index across 3,000+ public skills and private workspaces to tame duplication and security risk
- Automated Loops: Built-in loops for code review, skill optimization, security, docs, evaluation and cost benchmarking
- Works with any coding agent: Integrates Claude Code, Cursor, Codex, Antigravity and more
Use Cases
- Turn team coding standards into reusable rules
- Gate rules a linter cannot check in pre-commit and CI
- Tame duplicated and risky agent skills across an org
- Automate PR review against your own standards
- Evaluate model and agent combos on real tasks
Pros
- Standards owned by your team, versioned in the repo and reusable
- Verifiers fill the gap between linters and code review
- Write once, reuse locally, in pre-commit, and in CI
- Works with any coding agent, no vendor lock-in
- Free tier covers the core workflow
Pricing
Free tier at $0; Team at $100/month (5,000 credits); Enterprise custom. The default LLM judge is GPT-6 Luna, with an optional Jev model to lower cost and latency.
Summary
Tessl moves your team's standards out of people's heads and into the repo. Verifiers run an LLM judge over every change as a pass-or-fail gate for rules a linter cannot check; paired with a skill Registry and automated Loops, it is the governance layer that turns coding agents from writing code into shipping code against your standards.
Version History
- Releases Verifiers (2026-10-15): Releases Verifiers into general availability: rules written as JSON in the repo, scored pass or fail by an LLM judge on every change, reusable locally, in pre-commit and CI; targets team standards a linter cannot express, such as user-facing errors must tell the user what to do next and handlers must not log tokens; default judge is GPT-6 Luna with an optional Jev model.