AI Reviewer on the GAMMA-Ray Spectrum: A1 · R2 · C– · O2

PrompTugal.news draft · from the GAMMA-Ray report · beta · 2026-11-10

The GAMMA-Ray studio reads Codacy's AI Reviewer from its public documentation: it reviews GitHub pull requests when an admin turns it on, a person applies what it suggests, and its data use and limits are written down.

Codacy, founded in 2012 by Jaime Jorge and João Caxaria and headquartered in Lisbon, Portugal, runs code quality, security and coverage checks on repositories. Its AI Reviewer reviews pull requests on GitHub. A Codacy case study names BFI Finance, which switched off its previous AI review tool for it. We read it only from public sources, captured and hashed on 10 November 2026, and the published rules computed the card.

Autonomy 1, Assists (available). In the Manually setting a person starts each review, from the pull request or through the API. The reviewer posts a verdict, a summary and inline comments with suggested fixes; a person applies them.

Reach 2, Integrated (available). With the reviewer on, it combines GitHub pull request data, Codacy findings and a linked Jira or Linear ticket.

Control: not yet evidenced. Teams can write a review.md file that tells the reviewer what to ignore, but no page says the reviewer is bound by it. A page on how people approve, override or stop the reviewer would reach level 1.

Openness 2, Documented. The docs give the reviewer's size limit and the context it sends to the model, and say secrets are redacted. Published evaluation results with a method would reach level 3.

The card is beta and Codacy has not seen it. It can claim the card and add evidence; corrections are dated.

From public sources only; Codacy has not seen it. The levels come from [the report](./index.md) and its cards, which are corrected in place. Quote the card by linking to it.

O AI Reviewer da Codacy no GAMMA-Ray Spectrum: A1 · R2 · C– · O2

O estúdio GAMMA-Ray lê o AI Reviewer da Codacy a partir da sua documentação pública: revê pull requests no GitHub quando um administrador o ativa, e uma pessoa aplica o que sugere.

A Codacy, fundada em 2012 por Jaime Jorge e João Caxaria e sediada em Lisboa, verifica a qualidade, a segurança e a cobertura do código em repositórios. O AI Reviewer revê pull requests no GitHub. Lemos apenas fontes públicas, capturadas e com hash a 10 de novembro de 2026.

Autonomia 1 (opcional). No modo manual, uma pessoa inicia cada revisão. Alcance 2 (opcional). Junta dados do pull request, resultados da Codacy e um ticket do Jira ou do Linear. Controlo: ainda sem evidência. Um ficheiro review.md diz ao revisor o que ignorar, mas nenhuma página diz que o revisor fica obrigado a segui-lo.

Abertura 2. A documentação indica o limite de tamanho e os dados enviados ao modelo, e diz que os segredos são ocultados. O cartão é beta e a Codacy ainda não o viu; pode reclamá-lo e acrescentar evidência, e as correções são datadas.