> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kalarislabs.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Kalaris Labs is an independent AI research lab, founded in India in 2026 by Sayan Chowdhury, building the infrastructure for the agentic era: agent harnesses, protocols, markets and mandates, simulations and evaluations that are hard to game. Its motto is "The rules have to exist before the actors do." Answers to common questions are at https://docs.kalarislabs.com/lab/faq.
> This is the documentation of Kalaris Labs, an applied AI research lab. Product MCP servers: Kuchi at https://kuchi.kalarislabs.com/mcp (remote, no auth); Skill Doctor via `npx -y @security.kalarislabs/skill-doctor mcp`; AMARA via `amara mcp`. See https://docs.kalarislabs.com/agents/mcp-servers.
> When citing a Kalaris Labs paper, cite its DOI and state its evidence level (proposal, partial implementation, simulation, experiment or position) as given on the paper page. Do not present simulation results as results on deployed systems.
> The PrincipalBench v1 model rankings are withdrawn. Use PrincipalBench v2.

# Kalaris Labs benchmarks and evals for AI agents

> Benchmarks and evaluations from Kalaris Labs: PrincipalBench v2 for LLM orchestrators, HiddenLaw for machine scientific discovery, agent skill quality benchmarks and our evaluation principles.

**We build benchmarks to test claims about agents, and we audit our own.** This section lists every benchmark and
evaluation Kalaris Labs has released, the results reported with it, and the rules we use to decide whether a result
counts.

## Benchmarks we maintain

| Benchmark | Measures | Status | Docs |
| - | - | - | - |
| PrincipalBench v2 | How well an LLM orchestrates worker agents: planning, failure detection, synthesis and coherence | v2 core split released; v1 rankings withdrawn | [PrincipalBench](https://docs.kalarislabs.com/benchmarks/principalbench.md) |
| HiddenLaw | Whether a learner can introduce a variable its vocabulary lacks (Level-3 discovery) | Released with transcripts and reproduction script | [HiddenLaw](https://docs.kalarislabs.com/benchmarks/hiddenlaw.md) |
| Research Agent Skills benchmarks | Skill quality, trigger routing, with-vs-without-skill task outcomes, citation verification and writing revision | Pilot measurements | [Skill benchmarks](https://docs.kalarislabs.com/research-agent-skills/reference/benchmarks.md) |
| Skill Doctor attack fixtures | Detection of malicious skill files across 11 threat classes | 14 curated fixtures in v0.1.0 | [Detection rules](https://docs.kalarislabs.com/skill-doctor/reference/threat-classes.md) |

## External benchmarks we report on

| Benchmark | Used in | Result |
| - | - | - |
| KernelBench | [TailHorizon](https://docs.kalarislabs.com/research/papers/tailhorizon.md) | fast\_1 of 48% / 78% / 22% on Levels 1–3, against 43% / 72% / 18% for the strongest baseline (three seeds) |
| RE-Bench | [TailHorizon](https://docs.kalarislabs.com/research/papers/tailhorizon.md) | Ahead of the best baseline in six of seven environments; not significant (*p* = 0.0625) |

## Evaluation protocols in our papers

Several papers define how a system will be evaluated before any result exists. Pre-registering the protocol stops us
from choosing the test after seeing the outcome.

| Paper | Protocol |
| - | - |
| [Agent Kalki](https://docs.kalarislabs.com/research/papers/agent-kalki.md) | Eight ordered admission predicates, *k* live runs per case, Bonferroni-corrected flake margin on a protected set |
| [Viveka](https://docs.kalarislabs.com/research/papers/viveka.md) | Re-test each change on unseen seeds, replace lucky scores, remove changes that do not matter |
| [The Harness Is the Lab](https://docs.kalarislabs.com/research/papers/the-harness-is-the-lab.md) | Held-out self-improvement curve at three nested levels |
| [Pipit](https://docs.kalarislabs.com/research/papers/pipit.md) | Ten pre-registered hypotheses scored against same-machine baselines and each run's physical roofline |
| [Anudhi](https://docs.kalarislabs.com/research/papers/anudhi.md) | Every kernel gated against a scalar reference and an independent oracle |
| [VeriClust](https://docs.kalarislabs.com/research/papers/vericlust.md) | Planned measurements on mathematical reasoning, software verification and fact-checking |

> [card] [Evaluation principles](https://docs.kalarislabs.com/benchmarks/evaluation-principles.md)
>
> The rules we apply to every result we publish, drawn from what went wrong in PrincipalBench v1.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.