do

name	do
description	Manifest executor. Works through Deliverables verifying every Acceptance Criterion and Global Invariant. Use when executing a manifest, running a plan, implementing a defined task, or when the user asks to run, execute, implement, or ship a manifest-backed plan.
argument-hint	<manifest-path>
user-invocable	true

Work toward the manifest's Deliverables. Before calling /done, verify every Acceptance Criterion and Global Invariant by spawning one subagent per criterion using its verify.prompt: verbatim — no rewording. (Multi-repo manifests declaring Repos: prepend the path map per define/references/MULTI_REPO.md; otherwise nothing wraps the author's prompt.) Every verifier is a general-purpose subagent driven by verify.prompt — when a gate needs specialized behavior the prompt activates a skill (e.g. the review-code skill with a dimension, or the check-pr skill). There is no verify.agent field. The optional verify.model: selects the model. Respect phase: ordering — serial across phases, parallel within. Each verifier returns PASS, FAIL, or BLOCKED; all must PASS before /done. Any BLOCKED routes via /escalate. FAIL bodies carry findings or a natural-language hint — read them and act on what they say.

Code-change fix attempts iterate until they pass or /do judges them genuinely unrecoverable → /escalate. Other retry shapes (waiting, retriggering, replying with or without resolving, mechanical syncs) aren't fix attempts — they follow the verifier's findings directly. When a hint or finding indicates terminal / unrecoverable / human-decision-needed, route to /escalate — autonomously amending the manifest to suppress the block is forbidden.

When a finding carries external review input — a PR review comment or bot suggestion, as opposed to the manifest's own Acceptance Criteria and Global Invariants, which stay authority and must be satisfied — judge it before acting instead of implementing it to make the thread go away. Weigh whether it's correct, whether it serves this PR's intent, and whether addressing it is proportionate to that intent — a valid point that needs work beyond the PR's intent belongs in separate work, not this PR. Adopt the comments that clear that bar; on the ones that don't — a false positive, or a valid-but-separate-scope ask — reply with your reasoning rather than changing code. Push back even on a human reviewer when you are confident, with a respectful reply that leaves the thread open for them to resolve; when you are not — a borderline-valid point, or a substantive design objection — surface it to the user, or in an autonomous run reply non-committally and leave it for a human rather than bulldozing. Making this call autonomously is safe because /do drives to mergeable, never merged — a human still reviews the diff before the button. When the user does want a beyond-intent ask incorporated, that is an amendment: invoke define again with the manifest path and the amendment context — /define reads "manifest path in args = amend" and applies targeted changes.

Caller overlays may narrow retry cadence without changing the manifest. In CI one-shot / no-wait contexts, execute immediately actionable findings (fix, test, commit/push when authorized, retrigger, reply, resolve, sync), then stop instead of executing long wait directives such as bash sleep <N>; reinvoke. If only wait-shaped findings remain, report the waiting state as pending; do not call /done, do not call /escalate, and do not keep the runner alive.

When a caller supplies a journal/log path (e.g. /babysit-pr --log), read it before deciding retries and comment judgments, and append to it after acting. Record what GitHub state won't reconstruct on its own — dead-end memory: fixes tried and reverted, approaches considered and rejected that left no commit — plus operational notes (retriggers, waits), so those decisions survive context compaction in a long-lived run. The journal is also where runaway protection lives: lifecycle verifiers like check-pr are stateless and report current state without counting cycles, so you own the stop condition — when the journal plus your run memory show a fix or wait has been retried well past the point of progress, route to /escalate (or, in no-wait mode, a pending summary) rather than looping. With no journal path supplied, behave exactly as without one — the journal is an aid, not a precondition.

Mid-/do user messages default to invoking define for amendment — the manifest is the source of truth, silent scope drift is worse than an extra amendment cycle. Pure questions about the manifest or process are answered inline.

Unattended launch. When /do is the top-level execution entrypoint, establish a durable goal-setting backstop whose completion contract is auditable from the transcript after /do reads the manifest. The contract must require a gate ledger covering every Acceptance Criterion and Global Invariant: gate id, phase, verify.prompt source, latest independent verifier verdict, evidence, and freshness relative to the last relevant implementation change. Completion requires every listed gate to have fresh PASS evidence and /done reported. Unverified, FAIL, stale, BLOCKED/actionable, or escalation-pending gates are non-terminal; if implementation changes after a PASS, mark affected gates stale and re-verify before /done. Do not accept self-attestation, "looks done", or summary claims in place of verifier output. If a broader parent workflow backstop is already visible (for example /auto's full-chain contract or /babysit-pr's PR-tend contract) and it carries this manifest gate-ledger condition, do not set or print a second narrower goal; operate under the parent contract. If the visible parent lacks that condition, supplement it by setting or printing the manifest-completion contract before continuing. Otherwise, if the active harness exposes a goal-setting or continuation capability, set the manifest-completion contract directly; if not, print the copy-pasteable contract for the user to apply manually. The contract should also carry the objective, constraints, stop/block condition, and compact progress expectations when useful, so continuation keeps running until all gates pass rather than stopping at the first turn's end.

Input. <manifest-path> — required; no args → halt with usage. Read the manifest fully before any execution. Multi-repo manifests (declare Repos: [name: path, ...] in Intent) — use absolute paths in tool calls when working in a non-cwd repo.