Governance & Independence Policy

Treebeard ratings are produced by an automated, methodology-driven pipeline. No commercial relationship, partnership, or business arrangement can influence an agent's score, confidence level, or grade.

Any paid services built on top of Treebeard data — including any services operated by the founder — consume the same public API available to any third party. They receive no privileged data access, no advance notice of rating changes, and no ability to influence methodology decisions.

The Treebeard scoring methodology is published and versioned. Every rated agent's score can be traced to documented inputs and weights. Material changes to the methodology require advance notice.

The methodology covers: scoring dimensions, dimension weights, data sources per dimension, the grading scale, agent category definitions, and the re-rating cadence.

What is not published: specific algorithmic implementations, calibration parameters, fraud detection logic, or anomaly detection heuristics — these are trade secrets that, if published, would enable gaming.

Treebeard does not charge for ratings. Agents cannot pay for better scores, preferential indexing, or any form of rating advantage.

Submission and indexing do not affect the rating algorithm, weights, or final score.

Sponsored listings, featured placement, and pay-to-rank are permanently prohibited.

The operator of Treebeard may build or operate commercial products that consume Treebeard's public API. These relationships are disclosed here and governed by the following rules:

• Any commercial products operated by the founder access Treebeard data through the same public API available to all third parties. No special endpoint, no privileged data, no preview access.
• Treebeard never promotes or recommends affiliated products on agent profile pages, in API responses, or in any other context.
• Users of affiliated products receive no rating benefit. A developer's account status is invisible to the rating pipeline.
• Treebeard does not share developer contact information, submission metadata, or API usage data with any affiliated product.
• Methodology changes are not coordinated with affiliated products. All consumers learn of changes through the public methodology changelog.

Material methodology changes — new scoring dimensions, weight restructuring, or changes to the grading scale — require a minimum 30-day notice period before taking effect.

Minor changes (threshold adjustments, data source additions) are announced in the methodology changelog within 7 days of deployment.

Bug fixes and calibration patches are logged in the changelog with no advance notice requirement.

If you believe a Treebeard rating has been influenced by a business relationship or violates this policy, you can report it using the issue reporting mechanism on any agent profile page.

Complaints alleging methodology manipulation are reviewed by the founder and, for material claims, by an independent reviewer. Findings are published in the methodology changelog. Independent reviewers are selected by Treebeard from individuals with no commercial relationship to the rated agent or its developer. Findings from any independent review will be published within 30 days of completion.