Skip to content

Reviews

DevEx › Audit Chain › Reviews (?tab=reviews; its own DevEx row until 2026-10-31) shows the findings per review type. Every pull request in the org is reviewed on the platform’s own model, and no review finding is left as a bare comment. This page describes the ladder a finding goes down until a clean review clears it.

The ladder

Review

Runs on pull_request opened/synchronize, on a PR comment by an OWNER, MEMBER or COLLABORATOR that contains @tide_review, @tide_check, @code_review or @review, on workflow_dispatch -f pr=, and on repository_dispatch: tide-review. Bot comments never trigger it.

Runink TIDE first answers with a ✔️ comment and the tide-review label, then agents/reviewer posts PR review comments on the coder tier (REVIEW_MODEL: coder). Reviews are serialised by the model plane, not by a GitHub concurrency group: the plane serves one coder call at a time, and a review’s clock starts only once it is served, so a queued review is never cancelled. A review that could not run fails its check and says why on the PR. If any findings survive, the PR gets the label review:findings-open. Only a clean review clears it.

Fix (agents/fixer, inline)

The review job hands every finding, of any severity, to the fixer inline. The fixer makes a one-shot patch and opens a draft correction PR onto the reviewed branch through the GitHub REST/Git Data API. It never merges. The plan gate judges the patch first; a dissent means nothing is pushed. You can also run it by hand with an @tide_fix PR comment.

Agent drafts (heads fixer/* and tide-resolve/*) are not fixed in place. Their findings go straight to the resolver.

Resolve

When the fixer opens no fix, or the PR is itself an agent draft, the review hands the PR to the resolver. The resolver runs a full tide session start --base <PR head> with TIDE_SESSION_VERIFY=build. It raises that to test only when a live bwrap probe passes on the runner. It opens a draft PR onto the PR’s own branch and closes the agent drafts it supersedes.

It resolves on the root human PR of a chain, with every chain PR’s findings in the goal. A failed attempt starts again with a bigger budget, up to three attempts in a row:

Attempt--max-turns--timeout
11240m
21870m
3+24100m

After that it stops. A new review (a push) starts the ladder again, and a workflow_dispatch with the repo and PR runs it by hand. (The 3-hourly sweep was removed on 2026-11-07.)

Merging a draft onto the PR branch re-triggers the review, and a clean review clears the label. “Resolved” therefore means the reviewer agrees, not that an agent said so.

repository_dispatch, not workflow_dispatch, is used throughout because the runink GitHub App has actions: read only.

Repos without their own caller: the central sweep

A repository does not need its own review workflow. A central sweep runs every 10 minutes. It lists open, non-draft PRs org-wide that were updated in the last 7 days, and for each head SHA nobody has reviewed:

  1. sets commit status core/review = pending on the head SHA;
  2. starts the review against that target, with an App token scoped to it.

A dispatched review reports through the core/review status instead of a review check-run, because the App has statuses: write but only checks: read. A core/review status still pending after 90 minutes is started again. Nothing is configured per repo; a repo created tomorrow is covered.

The CI counterpart for FORGE and forged repos is central verification, which reports core/ci. See Central verification.

Comment tags

The tag-help agent answers any comment that looks aimed at TIDE (@core… or tide @…) but carries no recognised tag, by listing the tags that work:

TagWhat it starts
@tide_review (also @tide_check, @code_review, @review)sovereign code review of this PR
@tide_fixa draft correction PR for this PR’s findings
@tide_deploythe deployer agent on a task issue
@tide_ship [images]build image(s) and roll what runs them (default tide-operator)
tide @builda quick build and vet check
tide @curatethe curator
@tide_datagovassess the connected data sources
@tide_judgementjudge the findings submitted to TIDE
@tide_reconreconcile the rule book against the engine

@tide_review and @tide_fix only work on a pull request; on an issue there is no diff to read.

@tide_ship and @tide_deploy are different commands: @tide_deploy belongs to the deployer.

Where you see it in the console

DevEx › Reviews (?tab=reviews, GET /api/reviews) shows, per review type, what is running, the history, and the findings. For type code, runs come from the run store (reviewer, fixer, resolver) and findings from open PRs labelled review:findings-open.

TypeRuns fromFindings from
codereviewer, fixer, resolveropen PRs labelled review:findings-open
riskriskthe “🛡️ Dependency risk report” issue
compliancecompliancethe compliance findings and the “📋 Compliance status” issue
dataQualitydatagov and CapEx scan runsthe data governance findings and the CapEx findings
dataConsistencyCapEx scan runsthe CapEx consistency dimension
aiMaturity, dataMaturitynone, computed on readthe maturity assessment

A run whose findings were not recorded says so, never a zero count that would read as a clean review. A run’s detail adds what the agents posted on GitHub.