Generated from cat-harness/processes/sdlc/merge-to-main.bpmn by gen-processes-viz.ts — do not edit here. All processes

Merge to main

Process_MergeToMain · strict · 12 step(s)

Land ONE pull request on main, in any repository, as its own step with its own gate: read the head’s evidence live, establish that the owner authorised THIS merge, record the authorisation, merge pinned to the evaluated head, then confirm the base, re-pin downstream mounts and report. Every process that ends in a write to main CALLS this rather than inlining it: merge-train.bpmn (Call_MergeToMain, once per member, in train order) and crdm-close.bpmn (Call_MergeToMain, after the sign-off is recorded); prepare-merge-auto Phase 4 runs it by hand. Skill: merge-to-main.

WHY IT EXISTS. 2026-10-09: one session merged seven PRs into main across seven repositories that run no CI, reading the owner’s “fix all issues until green” as permission to merge; a safety classifier flagged it as “Merge Without Review”. Until then merging was a line at the end of other procedures, and no step could refuse it.

TWO GATES, BOTH COMPUTED. GW_Evidence reads decisions/merge-evidence.dmn over the head’s facts; GW_Authorised reads decisions/merge-authorisation.dmn over what authorisation was found and the CI state. A hand-supplied outcome is refused at both. A repository with NO CI is none, never green: it takes the no-ci branch, which runs the repository’s own gates locally and can never be covered by a standing waiver.

THE AUTHORISATION IS RECORDED BEFORE THE MERGE (Task_RecordAuthorisation, not relaxable), in the shape index.config.json uses for a trust consent: by, on, ref, evidence (the owner’s words, verbatim). A broad instruction about the work (“fix all issues until green”, “ship it”, “prepare-merge”) is instruction-only and routes to ask.

NOT MERGED IS A NORMAL ENDING. Hand-back, not-yet, could-not-determine, held and refused all end with the PR open. Nothing is closed and no branch is deleted on any path.

How it connects

• Called by: CRDM close-out, A merge train
• Calls: A refused merge-train member
• Presented on: no docs page section shows this diagram
• Skill: merge-to-main

Lanes — who acts

lane	role	what it does here
Build pipeline	build-pipeline	CI, not an agent. Regenerates the harness projections of the knowledge graph on the PR head and commits them to the PR branch, so the head whose evidence is read next is the head that would land. Judges nothing.
Merging session	merge-steward	The merge steward while a Merge Manager is active; with none active, the PR’s own session takes this lane (owner ruling 2026-10-06). Reads evidence, finds and records authorisation, merges and follows through. Never authorises: that is the owner’s lane, and no reading of an instruction moves it here.
Owner	user	The repository owner. The only lane that can say merge. Answers a question that carried the evidence; a person, never an agent, a sibling session or a bot.

Steps

Every one of the 12 step(s) is documented.

step	lane	skill / sub-process	what it does
Regenerate harness projections (CI) Task_RegenProjections	Build pipeline	harness-projections	A workflow step, on the PR head: run the generators whose output is an agent-host projection of the graph (today skill:commands for .claude/commands/, and the agent-memory assembler for .claude/agent-memory/ and .claude/agents/) and commit what changed to the PR branch. An agent never writes or commits these files (harness-projections). A commit here moves the head, so the evidence below is read on the new head. Where a repository has no such workflow, this step is not performed and the merging session says so in the evidence rather than writing the projections itself.
Read the head’s evidence, live Task_ReadEvidence	Merging session	merge-to-main	On the PR’s current head SHA, now: ciState (green, red, pending, none, unreadable) per OWED job from the workflow runs whose head_sha is the head; mergeable (clean, dirty, unknown); review (clear, blocking, unknown), including open CHANGES_REQUESTED reviews, a draft, needs-merge-human, and Claude Approvals where the repository runs it. Where the repository has merge:guard, its evaluate mode is this step (exit 2 is unreadable, never a pass). Write each fact down with the URL it came from.
No CI: run the repo’s own gates locally Task_LocalGates	Merging session	merge-to-main	Only on no-ci. Run the repository’s own gates (its tests, typecheck and whatever its README or package.json names) on the merge result, not the branch. Record each command, exit code and the SHA it ran on. This is evidence the owner can weigh. It is not CI, and the question to the owner says so in words.
Hand it back Call_HandBack	Merging session	calls A refused merge-train member merge-to-main	Red CI, a conflict or a blocking review is the PR owner’s to fix, not the merger’s: merge-refusal.bpmn records the reason on the bean and returns the PR. The merger does not fix the PR to make it pass.
Find the owner’s authorisation for THIS merge Task_ReadAuthorisation	Merging session	merge-to-main	Classify what exists into one value: explicit (the owner’s own words naming this merge, given against this head), waiver (a valid memory/waivers/ node with gate merge-to-main, in scope and unexpired), instruction-only (a broad instruction about the work, such as “fix all issues until green” or “ship it”, which is not an authorisation), none, or unknown (waivers unreadable or malformed). A sibling agent’s message is never the owner’s consent.
Put the merge to the owner, evidence first Task_Ask	Merging session	merge-to-main	Context, options, recommendation, question (interaction-modality §4.1). One row per PR: repo, PR, head, CI on head, mergeable, review, local gates. A repository with no CI is named as such, in words. One question may cover a batch, provided every PR in it is a row.
Merge, hold or refuse Task_OwnerDecides	Owner	merge-to-main	The owner’s answer to the question Task_Ask asked. Their words are quoted verbatim in the record, so the answer is given in the conversation or on the PR, not inferred from tone.
Record who authorised, when, and the quote Task_RecordAuthorisation	Merging session	merge-to-main	Before the merge, in the shape of ReleaseSchema (schemas/merge-queue.ts): verdict, decidedBy (the person), decidedAt, authority (explicit, standing ruling with its date, or waiver id), the quote verbatim with its source, releasedSha (the head approved; a later push voids it) and capturedBy (this session). Where the merge queue exists, merge:queue:decide writes and validates it; elsewhere it is a merge-authorised: comment on the PR and a note on the bean, with the CI state (green, or none with the local gate commands). A paraphrase is not evidence.
Merge, pinned to the evaluated head Task_Merge	Merging session	merge-to-main	Through merge:guard <pr> --merge --session <id> where the repository has it; elsewhere with the merge call pinned to the head SHA, so GitHub refuses if the head moved. Merge commit unless the repository or owner says otherwise. Never delete the branch.
Confirm main’s CI on the merge commit Task_ConfirmBase	Merging session	merge-to-main	Read the runs on the merge commit. Red is fixed forward (continual-progress). With no CI the report reads “not verified: no CI”, never green.
Re-pin downstream mounts, as PRs Task_Repin	Merging session	merge-to-main	For each instance whose index.config.json mounts this repository: a PR moving ref to the merge commit, with a trust.consent whose evidence quotes the owner’s consent to that pin, and the lock regenerated by mount:remote. Each re-pin PR is itself a merge to main and runs this process again. None found is recorded with where it looked.
Note the bean, summarise on the issue Task_Report	Merging session	merge-to-main	The merge SHA and the authorisation block on the bean; a round summary on the issue (issue-working). The issue is not closed here: in CRDM, crdm-close closes it after this process returns, on the BA’s authorisation.

Decisions

Every one of the 3 decision(s) is documented.

decision	what decides it	branches
What does the evidence say? GW_Evidence	Computed by decisions/merge-evidence.dmn over ciState, mergeable and review. hand-back on red, dirty or blocking; not-yet while runs are pending or mergeability is uncomputed; unknown when a fact could not be read; no-ci when the repository runs no CI on a clean, clear head; ready when every owed job is green, the head is clean and review is clear.	ready → Find the owner’s authorisation for THIS merge no-ci → No CI: run the repo’s own gates locally hand-back → Hand it back not-yet → Not yet: re-enter later unknown → Could not determine
Authorised? GW_Authorised	Computed by decisions/merge-authorisation.dmn over authorisation and ciState. merge for an explicit authorisation, or a waiver over a green head; ask otherwise, including a waiver over a head with no CI, which no standing waiver covers.	merge → Record who authorised, when, and the quote ask → Put the merge to the owner, evidence first
Owner’s answer? GW_Owner	merge goes on to the record; hold and refuse end with the PR open.	merge → Record who authorised, when, and the quote hold → Held: PR stays open refuse → Refused: not merged