Editorial and verification policy

Scope and independence

EcomAgentTools is an independent discovery, review, and comparison platform for ecommerce AI Tools, Skills, Prompts, and operating guides. We evaluate whether a resource helps with a defined ecommerce task, not how long its feature list is.

A vendor, author, maintainer, or contributor cannot buy a score, ranking, verification label, review outcome, or correction decision. Sponsored or affiliate relationships must be clearly disclosed where they exist and cannot replace the evidence rules on this page.

Evidence hierarchy and source labels

Product identity, features, pricing, platforms, limits, and installation facts should come first from the official product site, documentation, pricing page, repository, app marketplace, or another primary source. A reviewer must be able to open the source and connect it to the statement.

We distinguish vendor claims, public platform metrics, community reports, editorial inference, and hands-on testing. We do not label a result as hands-on tested without an actual recorded test, and we do not present marketing language as an independent finding.

What verification means

Verified means that EcomAgentTools checked the resource identity, official or primary source, and the relevant public facts during the review workflow. Verification is attached to a review record; it is not automatically granted when a URL is parsed or a draft is saved.

Products change. A verified page can still become outdated after review, so material facts should identify a source and freshness period where the data supports one. Unresolved, conflicting, or unavailable facts should be omitted or described as unconfirmed instead of guessed.

Editorial scores

Tool scores use capability, time to value, price to value, and market presence, plus community ratings when enough ratings exist. Skill scores use capability, documentation, reproducibility, and ecommerce fit, plus community ratings. Prompt pages do not receive an editorial score.

Community weight grows with the number of ratings. For Tools it can reach 30% of the combined score; for Skills it can reach 35%. With no community ratings, that weight is redistributed across the remaining axes. The score explanation on each page shows the active axes, values, and weights.

A score is a structured editorial aid, not a universal verdict. Market data, product documentation, implementation difficulty, and pricing can change, so readers should use the score together with the page’s evidence, limits, and selection guidance.

Community ratings and reviews

Community ratings are collected separately from editorial assessment. Reviews can score the dimensions shown in the review form and should explain the task, setup, outcome, and constraints. Aggregate community results appear only from eligible reviews in the moderation system.

We may reject, archive, or remove spam, copied reviews, conflicts that are not disclosed, rating manipulation, personal attacks, or claims that cannot be connected to the described experience. A community score does not override factual errors in a listing.

Pricing, value, and comparisons

A public paid product should show the current price and billing unit when an official source provides them. Price cards stay concise; order, contact, seat, ticket, usage, plan limits, and total-cost considerations belong in the pricing analysis when they affect value.

Pricing analysis should compare the plans in plain language and identify the strongest value for a realistic user profile. It should explain important thresholds and add-on costs without inventing a price. “Contact sales” is used when a reliable public amount is unavailable.

Comparisons are based on the same target task and should explain differences in platform support, operating limits, setup, control, and cost. More features alone do not make a product the better choice.

Authorship and attribution

A source link does not prove that a person shown on the source page created the listed Prompt or Skill. We name an author, developer, provider, or maintainer only when primary evidence supports that relationship.

When authorship cannot be confirmed, the page describes the resource as compiled, cataloged, or provided through a platform rather than assigning unsupported credit. Ownership and management claims follow a separate audited approval process.

AI assistance and human review

Automated systems may classify submitted URLs, extract public facts, draft descriptions, translate content, and identify possible images or missing fields. Automated output is working material, not evidence by itself.

Reviewers are expected to compare material claims with accessible sources, separate fact from analysis, remove unsupported content, and request changes when necessary. The site must not claim an independent test, author identity, price, or product capability merely because an AI system generated it.

Corrections and updates

Readers, contributors, authors, and providers can report inaccurate facts, unsafe links, wrong attribution, missing disclosure, or outdated pricing from the relevant page. A report enters the review queue and does not edit the public page automatically.

A useful correction identifies the page, the exact statement, the requested change, and a primary source. EcomAgentTools may update, unpublish, archive, merge, or restore a record after review while preserving an audit trail.

Related policies and next steps

Continue with the account, data, content-review and community-contribution rules. If you find an error, report it from the relevant resource page and identify the exact statement that needs review.