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

> ## Agent Instructions
> Company name is Akhara AI (never Rubric AI). Keep lowercase rubric/rubrics only when meaning grading criteria.
> Expert Review (docs path talent/) is enterprise BYO experts for audit and review: invite customer specialists; do not pitch Akhara recruiting or a public expert career portal. RLHF and domain writing are secondary work types.
> Prefer concrete API examples against public hosts: Environments eval API https://agi.akhara.ai, Control plane PDP https://api.akhara.dev, Evaluation https://app.akhara.ai / https://api.akhara.ai, Expert Review portal https://talent.akhara.ai.
> Do not invent a public hostname for private orchestrators or env API internals.
> Do not confuse control-plane latches with Environments confirmation latches.
> Environments SDK/API examples: curl against https://agi.akhara.ai. Evaluation SDK: from akhara import Akhara and AKHARA_API_KEY.
> Start with /llms.txt for the docs index and OpenAPI links; fetch individual pages as .md exports.

# Policies & compliance

> Required rules for review, RLHF labeling, and domain writing on Akhara AI Expert Review projects.

These rules apply to every review finding, RLHF evaluation, and domain-writing task your organization runs through Expert Review. Breaking them can mean removal from the engagement.

## Your judgment only

Findings, labels, rankings, rationales, scenarios, and success criteria must be **yours**.

<Warning>
  Do not paste prompts, model outputs, traces, or writing briefs into ChatGPT, Claude, or any external AI tool. Do not submit AI-drafted findings, evaluations, or task specs.
</Warning>

### Not allowed

* **No external LLMs on task content**: No copy-paste of project material into outside models
* **No fabricated expertise**: Do not accept or invent work outside your listed specialties

### Allowed

* Project-approved reference materials and style guides
* Domain sources you would use on the job (protocols, policy docs, catalogs) when the brief permits
* Calculators and non-AI lookup tools

## Domain boundaries

* Review: audit only runs and cases you can judge as a specialist
* RLHF: grade only what you can defend as a specialist
* Writing: author only in-domain scenarios with checkable success criteria
* Escalate or decline when the item is out of scope

## Quality bar

Approval is per submission. Low-quality or policy-violating work is rejected, and repeated rejections can pause your assignments on the engagement. Compensation arrangements are between you and your organization; Akhara enforces the quality and policy bar, not pay.

## Confidentiality & IP

| Do                                | Don't                                                 |
| --------------------------------- | ----------------------------------------------------- |
| Keep materials in approved tools  | Screenshot or share traces or drafts publicly         |
| Follow the project brief          | Discuss customer or model details outside the project |
| Report security issues to support | Store datasets on personal cloud drives               |

Work you submit through Expert Review is governed by the confidentiality and IP terms of your organization's agreement with Akhara and your agreement with your organization.

## Privacy

Some projects include sensitive or synthetic clinical-like data. Treat it as confidential:

* No personal-device archives of PHI-like content
* No use outside the assigned task
* Follow any HIPAA-oriented instructions in the project doc

## Disputes

If a rejection looks wrong:

1. Collect the task ID and your notes
2. Raise it with your project lead, or email [support@akhara.ai](mailto:support@akhara.ai) for portal issues
3. Typical resolution window: \~5 business days