ultragoal

Imported from Yeachan-Heo/gajae-code (packages/coding-agent/src/defaults/gjc/skills/ultragoal/SKILL.md). Install upstream with npx skills add Yeachan-Heo/gajae-code --skill ultragoal. Copyright stays with the author.

Ultragoal Workflow

Use when the user asks for ultragoal, create-goals, complete-goals, durable multi-goal planning, or sequential execution over GJC goal mode.

Purpose

ultragoal turns a brief into repo-native durable artifacts and then drives execution through the unified goal tool as a UX bridge only. goals.json is the canonical source of goal identity and state; ledger.jsonl is the canonical proof stream for checkpoints, receipts, blockers, steering, and reviews. The inline goal tool and goal-mode-request create-bridge exist only to keep the agent's interactive loop focused on the current aggregate or story objective. Completion is verified purely from durable goals.json plus fresh ledger.jsonl receipts, never from inline goal state. The agent, not the CLI or hooks, calls goal({"op":"complete"}) or goal({"op":"drop"}) after durable run completion or cleanup; CLI commands and hooks never mutate goal state.

• .gjc/_session-{sessionid}/ultragoal/brief.md
• .gjc/_session-{sessionid}/ultragoal/goals.json
• .gjc/_session-{sessionid}/ultragoal/ledger.jsonl (checkpoint and structured steering audit events)

Existing aggregate plans with the legacy enumerated objective are migrated to the stable pointer objective on read, persisted to goals.json, retained in gjcObjectiveAliases for already-active hidden goal reconciliation, and audited with an aggregate_objective_migrated ledger entry.

• Nudge budget setting — the per-story give-up budget (gjc.ultragoal.nudgeBudget, default 10, non-negative integer) is read through one shared resolver in this exact order (first valid value wins):
  1. project .gjc/config.yml
  2. user <agentDir>/config.yml (normally ~/.gjc/agent/config.yml, honoring GJC_CODING_AGENT_DIR/PI_CODING_AGENT_DIR; XDG applies only to categorized data/state/cache subdirs, never the workflow config path)
  3. built-in default config.yml uses the nested (schema) form - gjc: { ultragoal: { nudgeBudget } }. Project configuration beats user configuration. The reported source is the canonical path of the winning file, or default. config.yml is the ONLY settings surface: the legacy settings.json files (project and config-root) are retired: the config-root ~/.gjc/settings.json is migrated once into the default global agent config.yml and its source removed, while the project .gjc/settings.json is retained for non-workflow settings (only its workflow keys are migrated into project .gjc/config.yml and no longer read unless a migration target is absent - a migration that could not publish (e.g. a read-only .gjc) leaves the retained legacy value effective as the previously configured override until it can publish). Invalid optional settings files continue to the next layer or the default (tolerant).

Corrupt current-session state recovery

When ultragoal detects its own current-session state is corrupt, tampered, unreadable, or stale on resume, run gjc state clear --force --mode ultragoal before reseeding or restarting. Scope the clear to the current session via --session-id, the command payload, or GJC_SESSION_ID; it clears only ultragoal state for that session and never clears other skills or sessions.

Always-used command examples

Use these exact gjc ultragoal commands before spending tool calls rediscovering syntax:

gjc ultragoal status
gjc ultragoal status --json
gjc ultragoal create-goals --brief "<brief>"
gjc ultragoal create-goals --brief-file <path>
gjc ultragoal complete-goals
gjc ultragoal complete-goals --retry-failed
gjc ultragoal quality-gate source-hash --json
gjc ultragoal quality-gate validate --quality-gate-json <quality-gate-json-or-path> [--goal-id <id>] [--json]
gjc ultragoal checkpoint --goal-id <id> --status complete --evidence "<evidence>" --quality-gate-json <quality-gate-json-or-path>
gjc ultragoal checkpoint --goal-id <id> --status failed --evidence "<blocker/evidence>"
gjc ultragoal record-review-blockers --goal-id <id> --title "Resolve final review blockers" --objective "<blocker-resolution objective>" --evidence "<review findings>"

Use these exact goal-tool calls for the inline goal state:

goal({"op":"get"})
goal({"op":"create","objective":"<printed aggregate or per-story objective>"})
goal({"op":"complete"})
goal({"op":"drop"})
goal({"op":"resume"})

drop clears the active goal without exiting goal mode; resume reactivates a paused goal.

Create goals

1. Decide on the brief. To produce multiple stories, separate them with a reserved @goal: delimiter line; the title follows on the same line and the objective is everything beneath it until the next delimiter:

   Shared brief constraints / context go here (optional preamble).
   
   @goal: Parse the intake CSVs
   Ingest reviewer CSVs from the watch dir, validate headers, and reject
   malformed rows with a per-row reason. Objectives can span multiple lines
   and contain `code`, "quotes", or commands — no escaping needed.
   
   @goal: Normalize records
   Map raw rows onto the canonical schema and dedupe by record id.
   
   @goal: Export the audit report
   Emit an audit-ready report covering every accepted and rejected row.

   Delimiter contract:

   • A @goal line is a story boundary only when it starts at column 0 (no leading whitespace) and the character right after @goal is :, whitespace (space or tab), or end-of-line. So @goal: Title, @goal Title, and a bare @goal line all open a story.
   • @goalish, @goals:, @goal-foo, @goal.foo, @goal/foo, and any indented or mid-line @goal are ordinary objective text, not delimiters. To keep a literal @goal line inside an objective, indent it.
   • A title-only block (no body) uses the title as its objective. An empty title borrows the first body line as the title. A block with neither title nor body is rejected — create-goals errors instead of writing a placeholder goal.
   • Preamble (any text before the first @goal delimiter) is global context/constraints only; it is retained in the brief but is not turned into a goal. Every executable story needs its own @goal block.
   • With no @goal delimiter anywhere, the whole brief becomes a single goal G001 (unchanged legacy behavior).

   Stories become G001, G002, … in order.

2. Run one of:

   • gjc ultragoal create-goals --brief "<brief>"
   • gjc ultragoal create-goals --brief-file <path>
   • cat <brief> | gjc ultragoal create-goals --from-stdin
   • gjc ultragoal create-goals --gjc-goal-mode per-story --brief "<brief>" only when one GJC goal context per story is explicitly preferred

3. Inspect .gjc/_session-{sessionid}/ultragoal/goals.json and refine if needed.

Create-goals granularity: merge validation-coupled stories

Before splitting a brief into many thin stories, check whether the candidate stories are validation-coupled. Merge validation-coupled stories into one goal and fan out executor slices inside that goal instead of creating one goal per slice. Two stories are validation-coupled when they share any of:

• the same feature stack (one story's code cannot be meaningfully verified without the other's),
• the same acceptance surface,
• the same red-team surface, or
• the same final review boundary (they can only be signed off as a unit).

Fanning out executor slices inside a single merged goal keeps one review/QA boundary while preserving parallel implementation. When validation-coupled stories must stay as separate goals for scheduling reasons, use an aggregate-mode validation batch (below) so the coupled review happens once at the final member.

Complete goals

Loop until gjc ultragoal status reports all goals complete:

1. Run gjc ultragoal complete-goals.
2. Read the printed handoff.
3. Call goal({"op":"get"}).
4. If no active GJC goal exists, call goal({"op":"create","objective":"<printed payload objective>"}) with the printed payload. In aggregate mode, if the same aggregate objective is already active, continue the current GJC story without creating a new GJC goal. If goal({"op":"get"}) shows a stale dropped goal (status "dropped") and a new aggregate must start, no extra cleanup is needed — goal({"op":"create"}) succeeds directly. If a previous aggregate is still active and you genuinely need a fresh start in the same session, call goal({"op":"drop"}) first, then goal({"op":"create"}).
5. Complete the current GJC story only.
6. Run a completion audit against the story objective and real artifacts/tests.
7. Before any --status complete checkpoint, run the mandatory final cleanup/review gate below. In aggregate mode, do not call goal({"op":"complete"}) for intermediate stories; checkpoint each story while the aggregate objective is still active. On the final story, create the final aggregate receipt first; only after that receipt exists may goal({"op":"complete"}) run.
8. Checkpoint the durable ledger. Complete checkpoints require --quality-gate-json only: gjc ultragoal checkpoint --goal-id <id> --status complete --evidence "<evidence>" --quality-gate-json <quality-gate-json-or-path> A successful complete checkpoint is story completion, not automatic run completion. Read the checkpoint output: when it prints Next ultragoal goal: <id>, continue that active story under the same aggregate GJC goal; when it prints All ultragoal goals are complete, the durable run is terminal. gjc ultragoal complete-goals remains the supported manual next-story command if continuation output was missed.
9. If blocked or failed, checkpoint failure: gjc ultragoal checkpoint --goal-id <id> --status failed --evidence "<blocker/evidence>"
10. For legacy per-story completed-goal blockers, preserve the non-terminal blocker with: gjc ultragoal checkpoint --goal-id <id> --status blocked --evidence "<completed legacy GJC goal blocks goal create in this thread>"
11. Resume failed goals with gjc ultragoal complete-goals --retry-failed.

Blocker triage and pause discipline

An active Ultragoal run must not give up on a blocker by pausing the goal and asking the user. Classify every blocker before deciding what to do, and default to resolvable when unsure:

• resolvable — anything the agent can act on: failing tests, missing implementation, a dependency to install, an ambiguous-but-inferable detail, investigation. Never pause. Exhaust autonomous resolution first: investigate, gjc ultragoal steer --kind add_subgoal --title "Investigate blocker" --objective "..." --evidence "..." --rationale "...", delegate an executor, or preserve the blocker durably with gjc ultragoal checkpoint --status blocked / gjc ultragoal record-review-blockers and keep scheduling the next goal.
• human_blocked — only the user can act: credentials/secrets, a manual or physical step, an external approval/decision, access the agent lacks. Pause is the last resort and is gated.

goal({"op":"pause"}) is blocked at runtime while an Ultragoal run is active unless the latest blocker_classified ledger event is human_blocked and a later bound clean pause terminal critic verdict is recorded for it (see Terminal critic gate). assertUltragoalPauseAllowed first consumes a pre-existing give-up nudge (a durable ledger write) before it runs the read-only pause diagnostic; only isUltragoalPauseBlocked is a pure reader. To pause, first record the human-only classification and capture its event id, then record the terminal critic's clean bound pause verdict, and only then pause:

gjc ultragoal classify-blocker --classification human_blocked --evidence "<the specific human-only dependency>" [--goal-id <id>]
gjc ultragoal record-critic-verdict --terminus pause --classification-event-id <eventId> --verdict OKAY --evidence "<terminal critic evidence>"
goal({"op":"pause"})

Recording --classification resolvable is an audit note only; it never authorizes a pause. The ask tool stays blocked during active runs regardless of classification — record unresolved decisions as durable blockers instead of prompting.

Dynamic steering

Use gjc ultragoal steer when real findings or blockers prove the current story decomposition should change while the aggregate objective and constraints stay fixed. Steering is explicit-only and evidence-backed; broad natural-language requests are rejected instead of guessed.

Allowed mutation kinds are:

• add_subgoal
• split_subgoal
• reorder_pending
• revise_pending_wording
• annotate_ledger
• mark_blocked_superseded

Examples:

gjc ultragoal steer --kind add_subgoal --title "Investigate blocker" --objective "Validate the blocker and report evidence." --evidence "log/test output" --rationale "The blocker changes the safe execution order." --json
gjc ultragoal steer --kind split_subgoal --goal-id G002 --replacements-json '[{"title":"Fix parser","objective":"Resolve parser blocker."},{"title":"Verify parser","objective":"Run focused parser verification."}]' --evidence "Implementation split found two separable risks" --rationale "Splitting keeps each sub-goal independently verifiable." --json
gjc ultragoal steer --kind reorder_pending --order-json '["G003","G002"]' --evidence "Dependency order changed after investigation" --rationale "G003 must land before G002 can proceed safely." --json
gjc ultragoal steer --kind revise_pending_wording --goal-id G002 --title "Clarify blocker story" --evidence "The current title hides the actual blocker" --rationale "Clear wording keeps the ledger auditable." --json
gjc ultragoal steer --kind annotate_ledger --evidence "User changed release ordering at runtime" --rationale "The aggregate objective is unchanged, but the execution history needs an audit note." --json
gjc ultragoal steer --kind mark_blocked_superseded --goal-id G004 --evidence "The blocked work is no longer required because replacement evidence covers it" --rationale "No replacement sub-goal is needed; superseding only the blocked sub-goal unblocks final completion without changing the aggregate objective." --json

--directive-json and UserPromptSubmit structured steering are planned/deferred routing surfaces, not part of the native typed --kind CLI path described above.

Steering invariants:

• Do not edit the aggregate goal objective, original brief constraints, quality gates, or completion status. The aggregate objective is a stable pointer to .gjc/_session-{sessionid}/ultragoal/goals.json and .gjc/_session-{sessionid}/ultragoal/ledger.jsonl, not an enumeration of initial goal ids.
• Do not hard-delete goals, auto-complete work, weaken verification, or silently mutate .gjc/_session-{sessionid}/ultragoal.
• Accepted and rejected attempts append structured audit entries to .gjc/_session-{sessionid}/ultragoal/ledger.jsonl.
• Superseded goals remain in goals.json with steering metadata and are skipped for scheduling.
• Blocked goals without replacements are skipped for scheduling but still block final completion until later explicit steering replaces or supersedes them.

UserPromptSubmit structured steering directives are a planned/deferred routing surface. Normal prose does not mutate state.

Role-agent delegation guidance

Ultragoal execution should use GJC's bundled role-agent roster when a durable story is large enough to benefit from delegation:

• Use executor for bounded implementation, refactoring, and fix slices.
• Use planner for story sequencing or handoff refinement when execution uncovers a missing plan branch.
• Use architect for read-only architecture and code-review lanes, including CLEAR / WATCH / BLOCK status.
• Use critic for read-only plan or handoff critique before execution proceeds.

Implementation delegation guidance

Direct inline implementation by the leader is the default. Delegate to executor subagents only when the expected diffs land in genuinely different sub-domains, modules, or systems — separable surfaces with independent acceptance criteria and no shared-file contention. File count or line count alone does not force delegation; a large change confined to one domain/subsystem is usually better done inline or by a single sequenced executor.

Delegation is worth it when:

• The story spans multiple distinct sub-domains / modules / systems (e.g. a CLI surface plus an unrelated runtime subsystem plus docs tooling) whose slices can proceed in parallel without coordinating on the same files.
• Each slice can be bounded with explicit targets and acceptance criteria that are verifiable independently of the other slices.
• The leader's checkpoint/verification duties would otherwise be crowded out by juggling unrelated domains inline.

When delegating:

• Give each executor bounded targets and explicit acceptance criteria, and keep checkpoint/goal-state ownership in the leader.
• Parallelize only across genuinely different sub-domains/modules/systems; sequence anything with a real dependency or shared-surface overlap.
• Work within a single domain/subsystem stays with the leader as direct edits — do not split one cohesive change across subagents, and do not over-delegate trivial work.
• After integrating delegated slices, you MAY run architect / critic review lanes for early signal, but treat them as advisory: the canonical review is the boundary cohort gate below, and a slice-level lane never substitutes for it or its verdict. Skip slice review entirely when the boundary cohort will cover the same change set shortly. Worker agents never mutate .gjc/_session-{sessionid}/ultragoal or call goal tools.

When delegating with native subagents, an await timeout only limits the leader's wait. It is not subagent failure evidence and must not be used as a cancellation reason; inspect or continue independent work, and cancel only when the subagent has actually failed, gone off-track, or become unrecoverably wrong.

Subagent reuse and resumption (token efficiency)

Fresh spawns re-pay the full context ramp-up (file reads, domain orientation, contract restatement) on every delegation. When a later slice or lane targets the same sub-domain/module/system as a prior subagent of the same role, resume the prior subagent instead of freshly spawning:

• Track the subagent id per role + domain as it is created; on the next same-domain executor slice or same-scope architect review lane, resume that id and inject only the delta (new targets, new acceptance criteria, the updated frozen change set) rather than re-briefing from scratch.
• Reuse is domain-scoped: resume only when the prior context is an asset. A slice in a genuinely different sub-domain/module/system gets a fresh spawn — stale cross-domain context is a liability, not a saving.
• Resumability requires retained subagent resume metadata and a persistent parent session; use existing subagent resume/steer controls only. Route per attempt: running → steer/inject to the same id and await; queued → retain or await the same id; terminal (completed/failed/cancelled) with context available → resume the same id; context_unavailable, not_found, no_runner, or resume_failed → fresh spawn fallback for that slice.
• A resumed subagent is still the same worker under the same contract: it must not mutate .gjc/_session-{sessionid}/ultragoal, call goal tools, or absorb checkpoint/goal-state ownership, and review lanes (architect, critic) stay read-only when resumed.
• Resumption never weakens gates: a resumed architect review or executor QA lane must still evaluate the current frozen change set on its own evidence, not rubber-stamp its earlier verdict.

If an Ultragoal request has no approved plan or consensus artifact and the scope genuinely needs one, run ralplan first and preserve its PRD, test spec, role roster, and verification guidance in the Ultragoal ledger. Skip ralplan for small scope: work that fits a single reviewable PR and is tied to a single domain/subsystem can proceed directly from the brief — record that judgment in the ledger instead of running a planning round. Reach for ralplan when the scope spans multiple domains/subsystems, needs cross-cutting sequencing, or would not fit a single PR.

The Ultragoal leader owns .gjc/_session-{sessionid}/ultragoal/goals.json and .gjc/_session-{sessionid}/ultragoal/ledger.jsonl. Role agents return implementation/review evidence; they do not checkpoint Ultragoal or mutate goal state.

Native executor parallelism contract

Native subagent parallelism is a contract for bounded executor delegation, not a runtime scheduler:

• Use native executor parallelism only when a story's expected diffs fall in genuinely different sub-domains/modules/systems, each boundable by a per-slice coordination contract.
• Default to direct leader edits otherwise; sequence any work with real dependencies, shared-file overlap, or a single-domain footprint, and never parallelize work that lacks a safe contract.
• Worker agents MUST NOT mutate .gjc/_session-{sessionid}/ultragoal, call goal tools, make checkpoint decisions, own integration, or own final verification. The Ultragoal leader keeps those responsibilities.
• Workers must not run gjc ultragoal checkpoint: checkpoint authority stays with the leader after worker tasks are terminal. The leader checkpoints from worker evidence plus the current-session GJC goal snapshot, and performs no hidden goal mutation.

Before workers start, each per-slice coordination contract MUST name the target files/surfaces, independence assumptions, allowed coordination channel, conflict-escalation rule, expected evidence, and terminal status. Conflict or assignment changes remain leader-owned and must be auditable through durable ledger evidence.

For failed, timed-out, or contract-violating slices, record durable ledger evidence; preserve successful terminal slices only when safe; and reassign, retry, or collapse the invalid work to serial execution under an updated contract. Completion after parallel work still requires terminal worker evidence, leader integration, targeted verification, and the existing cleaner + architect + executor QA/red-team gate before checkpoint complete.

Boundary verification (aggregate default)

Heavyweight review runs once per boundary, not once per story. In aggregate mode the whole required-goal set is one implicit boundary by default: every checkpoint before the run's final required goal may present the lightweight deferredToBatch gate, and only the final goal carries the full strict gate. Nothing needs to be declared to get this — it is the default.

A deferred gate is just the proof the runtime cannot know: that targeted verification ran. Everything mechanical — kind, the batch tuple, deferredLanes, and the whole changeSet block (paths, changeSetHash) — is auto-filled from durable state and the computed cumulative git diff. Never hand-compute a hash. The minimal valid gate:

{
  "deferredToBatch": {
    "ranLanes": ["targetedVerification"],
    "targetedVerification": {
      "status": "passed",
      "commands": ["bun test <targeted suite>"],
      "evidence": "what was verified and how it passed"
    }
  }
}

deferredToBatch.ranLanes lists the lanes you actually ran (targetedVerification, plus optionally aiSlopCleaner / iteration); declaration and evidence must match in both directions. ranLanes can never claim architectReview or executorQa, and a deferred gate can never contain architectReview, executorQa, or validationBatchClose — review always belongs to the boundary, and deferring never manufactures approvals. Any optional field you do supply must match reality; a wrong value fails closed. Check with gjc ultragoal quality-gate validate before checkpointing.

Validation batches (explicit phase/module boundaries)

When one ledger is large enough that a single end-of-run boundary is too coarse, use an explicit validation batch to subdivide it into phase/module boundaries, each with its own final member. Validation batches are aggregate-only, explicit-only, and fail-closed. They are created only through --validation-batch-json; there is no inference from brief prose, no per-story batching, and no other batching input path.

Create a batch explicitly:

gjc ultragoal create-goals --brief-file <path> --validation-batch-json '[{"schemaVersion":1,"batchId":"VB001","memberIds":["G001","G002","G003"],"finalGoalId":"G003"}]'

Checkpoint contract summary — the full contract lives in the validation-batch-contracts fragment (skill-fragments/ultragoal/validation-batch-contracts.md); load it before checkpointing any batch member:

• Non-final members checkpoint complete with a single top-level deferredToBatch quality gate (kind validation-batch-deferred) proving targeted verification, a declaration-matched lane set, and a cumulative-since-base change set — never architectReview, executorQa, or validationBatchClose; deferring never manufactures fake review approvals.
• The final member (finalGoalId) checkpoints complete with the normal full strict gate PLUS a top-level validationBatchClose proof covering all members; out-of-order close is rejected, close state is append-only proof on the final member only, and batch invalidation is fail-closed. Like the deferred gate, every close field except coverageEvidence is auto-filled from durable receipts and the computed diff — the minimal close is {"validationBatchClose":{"coverageEvidence":"..."}} alongside the strict gate.

Intra-goal validation-lane parallelism

Cohort lanes are parallel by construction: the boundary gate freezes one sourceHash first, so cleaner, architect, and qa can run concurrently against the identical immutable snapshot and then join. Fall back to sequential lanes only when code is still changing (nothing can be frozen yet), when the red-team lane depends on architect fixes, or when architect findings gate the QA scope. Either way the lanes must join before checkpoint — no lane checkpoints independently, and repair work starts only after the join.

Internal Ultragoal sub-skill fragments

The completion-gate cleanup sweep is driven by ai-slop-cleaner, an internal Ultragoal sub-skill bundled as a kind: "skill-fragment" prompt with parent skill ultragoal (installed at skill-fragments/ultragoal/ai-slop-cleaner.md). It is analogous to deep-interview's auto-research fragment: loaded on demand for one specific hook, never a user-facing skill.

• It is not slash-command discoverable, has no public skill-listing entry, and is never resolvable through skill://.
• It is a read-only detector+reporter over the active story's changed files only: it never edits code, writes files, mutates .gjc/, checkpoints, calls goal tools, or spawns workflows.
• It classifies every finding as blocking or advisory across the full taxonomy (fallback-like masking vs. grounded, duplication, dead code, needless abstraction, boundary violations, UI/design slop, missing tests).
• The leader and a leader-spawned executor own all fixes; the cleaner reruns until zero blocking findings remain. Advisory findings live in the gate report only.
• Recursion guard: it must not spawn nested ralplan/deep-interview/ultragoal; broad or architectural findings are handed back to the leader as review blockers.

Boundary completion cohort gate

The heavyweight gate runs once per boundary generation, not once per story and not once per review pass. Intermediate stories use the lightweight deferred gate above; this section applies at the boundary (the run's final required goal, or an explicit batch's final member).

One generation freezes the change set and reviews it exactly once:

1. Run implementation verification for the boundary's cumulative change set.

2. Freeze the change set. Run gjc ultragoal quality-gate source-hash --json on the clean reviewed snapshot and use its sourceHash exactly. The runtime binds this digest to the integration base, merge base, normalized changed paths, captured diff, and untracked-content digest. Every lane in this generation inspects that same frozen snapshot; a lane verdict carrying a different sourceHash is rejected. Any later source or base change requires rerunning this command and starting a new generation.

3. Run the cohort lanes on the frozen snapshot — at most one cleaner, one architect, and one qa lane per generation. They may run in parallel because they share the frozen source; a second architect or QA lane in the same generation is rejected. The cleaner lane is the internal ai-slop-cleaner skill fragment run over the frozen change set: a read-only detector that emits an AI SLOP CLEANUP REPORT, and it still runs and records a passed/no-op report when there are no relevant edits. Its BLOCKING findings join the cohort findings rather than starting their own fix loop; advisory findings are included in the gate report only and are not written to the Ultragoal ledger.

4. Delegate an architect review covering all three lanes:

   • architecture-side: system boundaries, layering, data/control flow, operational risks.
   • product-side: user-visible behavior, acceptance criteria, edge cases, regressions.
   • code-side: maintainability, tests, integration points, and unsafe shortcuts.

5. Delegate an executor QA/red-team lane with typed executionMode: "ultragoal-red-team" (preferred) — or assignment text that explicitly labels Ultragoal completion QA/red-team — to build and run the e2e/red-teaming QA suite appropriate for the story. A bare executorQa field-name mention is not enough to activate the mode. This lane must try to break the change, not just confirm the happy path. It must start from the approved plan/spec/acceptance criteria, then user-facing contracts, and only then implementation code as supporting evidence. Plan/code mismatches are blockers, not items to paper over with implementation intent.

6. The executor QA/red-team lane must prove evidence by the real surface under test:

   • GUI/web surfaces require a valid automation transcript plus a non-uniform screenshot. Bare inlineEvidence text or typed receipts never prove live GUI/web execution.
   • CLI surfaces require a safe runtime argv replay (schemaVersion: 1, kind: "cli-replay", replaySafe: true) or the existing audited replayExempt path with a screenshot, automation, or PTY structural fallback. Runtime replay is limited to the pinned Bun runtime for bun --version or literal bun -e "console.log(...)"; the gate never executes model-authored test files. Shells, interpreters with code strings, path-qualified executables, package/git/network mutation commands, bun test, and arbitrary argv are rejected. Structured test-report fallback remains unsupported pending a separately reviewed provenance design.
   • Native/desktop/tui surfaces require a structurally valid screenshot, PTY capture with terminal control codes, or app-automation transcript.
   • API/package surfaces require a real artifact file or typed receipt whose artifact kind contains one of api, package, consumer, black-box, or test-report; examples: api-package-test-report, package-consumer-report, black-box-api-receipt. Algorithm/math surfaces require a real artifact file or typed receipt whose artifact kind contains one of property, boundary, edge, adversarial, failure, math, algorithm, or test-report; examples: property-test-report, algorithm-boundary-report. Bare inlineEvidence text alone is not sufficient for any surface.
   • The mandatory computer-use red-team suite (kill-switch-bypass, suspended-enforcement, permission-revoked, …) is conditional, not universal: require it only when computer/desktop control is genuinely part of the product surface being dogfooded. For every other product type, prove the change through the matching live surface instead — browser-use automation for web/GUI, bash/CLI live invocation or argv replay for CLI, and real artifacts or typed receipts for API/package/algorithm/math. Editing docs, prompts, or skills that merely mention computer-use does not by itself make the computer-use suite applicable; pick the red-team surface that matches what the change actually ships.
   • The runtime decides applicability from the change set, and it fails closed. Judgement about "what the change actually ships" does not override it, so check the paths before assuming the suite is skippable. gjc ultragoal checkpoint --status complete requires the suite whenever the computed change set touches computer source (crates/pi-natives/src/computer/**), the computer tool (packages/coding-agent/src/tools/computer.ts, packages/coding-agent/src/tools/computer/**), or a shared behavior registry — packages/coding-agent/src/config/settings-schema.ts, packages/coding-agent/src/tools/index.ts, packages/coding-agent/src/tools/renderers.ts. The registries are deliberately unconditional: they mix computer and non-computer entries, and a path-only or uninspectable change cannot prove computer controls were untouched, so any edit to them demands the suite even when the diff contains nothing computer-related. The suite is also required whenever change-set capture was incomplete. Generated bindings (packages/natives/native/index.{d.ts,js}), prompt/skill/doc files, and every other path do not trigger it on their own.
   • Practical consequence: a change that is not about computer-use at all — say a new settings key in settings-schema.ts — will still be gated on the seven mandatory cases. Do not fabricate them to get past the gate, and do not weaken the gate. Either supply a genuine suite, or treat it as a blocker and escalate to the operator (gjc ultragoal record-critic-gate-override exists for an authorized override).

7. The executor QA/red-team lane must report a matrix using executorQa.contractCoverage, executorQa.surfaceEvidence, executorQa.adversarialCases, and executorQa.artifactRefs. Not-applicable rows are allowed only in contractCoverage and surfaceEvidence; each status: "not_applicable" row requires contractRef plus reason. adversarialCases rows cannot be not-applicable.

8. Join before repairing. Fold all three lane verdicts and the final code review into the strict gate under iteration.reviewCohort (reviewGeneration, sourceHash, joined: true, and the three lanes). No lane may checkpoint on its own, and no fix work starts until the findings are joined. Clean means architectReview.architectureStatus, architectReview.productStatus, and architectReview.codeStatus are all "CLEAR", architectReview.recommendation is "APPROVE", executor QA statuses are "passed", iteration is "passed" with fullRerun: true, the cohort is joined with every lane clean and hash-bound, every evidence field is non-empty, every required matrix row is present, and every blockers array is empty. COMMENT, WATCH, REQUEST CHANGES, BLOCK, missing evidence, missing or shallow matrix rows, plan/code mismatches, or non-empty blockers are non-clean.

9. If the joined findings contain any blocker, do not checkpoint complete and do not call goal({"op":"complete"}). Record one consolidated blocker batch for all findings from the whole cohort instead of one story per lane:

   gjc ultragoal record-review-blockers --goal-id <id> --title "Resolve verification blockers" --objective "<blocker-resolution objective>" --evidence "<joined cohort findings>"

   Review-blocker recursion cap (#3613): record-review-blockers dedups identical-objective blockers (same trimmed objective + same blocked goal + open status) and bounds the number of unresolved review_blocker descents per blocked goal to 3. Descents 1..3 may exist; an attempt to create a 4th throws a typed review_blocker_recursion_cap terminal handoff (CLI exit 1, operator-visible marker) — never silently auto-completing findings. When the cap fires, record a human pause/escalation or resolve existing blockers before recording more.

10. One consolidated fix batch produces exactly one new generation. Re-freeze the fixed source as a new sourceHash, bump reviewGeneration, and set deltaOnly: true with priorGenerationSourceHash and the deltaPaths actually changed. Generation 2+ reviews are delta-only: they may not pull in unrelated scope without an explicit scopeExpansion carrying severity, novelty, and justification. Repeat until a generation joins clean.

11. Only after a generation joins clean, checkpoint the story as complete with a structured quality gate. The terminal critic runs once on that final joined generation; when criticReview.sourceHash is present it must match the cohort's sourceHash. The checkpoint creates a receipt in ledger.jsonl; goals.json.status alone is not proof. In aggregate mode, the final aggregate receipt must exist before the agent calls goal({"op":"complete"}) to reconcile the inline UX goal state.

While an Ultragoal run is active, the ask tool is blocked for all agents. Record unresolved review decisions as durable blockers with gjc ultragoal record-review-blockers instead of prompting interactively.

The native checkpoint --status complete command rejects missing or shallow gates, and reports all structural, evidence, surface, cohort, and declaration errors in one run rather than one per attempt. Each diagnostic carries a stable path, a stable machine-readable code, and a human message.

Validate before you checkpoint. gjc ultragoal quality-gate validate --quality-gate-json <json-or-path> [--goal-id <id>] [--json] applies exactly the same rules as checkpoint --status complete (including deferred-vs-boundary gate selection and artifact existence checks) but is strictly read-only: it never touches goals.json, ledger.jsonl, or goal state. It exits non-zero with the full diagnostics list when invalid, so authoring a gate is one pass instead of an edit/retry loop. --quality-gate-json must include:

{
  "architectReview": {
    "architectureStatus": "CLEAR",
    "productStatus": "CLEAR",
    "codeStatus": "CLEAR",
    "recommendation": "APPROVE",
    "evidence": "architect review synthesis across architecture/product/code",
    "commands": ["architect review command or agent evidence id"],
    "blockers": []
  },
  "executorQa": {
    "status": "passed",
    "e2eStatus": "passed",
    "redTeamStatus": "passed",
    "evidence": "executor-built e2e and red-team QA commands/results",
    "e2eCommands": ["bun test:e2e"],
    "redTeamCommands": ["bun test:red-team"],
    "artifactRefs": [
      { "id": "<ref-id>", "kind": "<surface-appropriate kind; see step 6>", "path": "artifacts/<file>", "description": "live-surface evidence" }
    ],
    "contractCoverage": [
      { "id": "<id>", "contractRef": "<approved contract id>", "obligation": "<required behavior>", "status": "covered", "surfaceEvidenceRefs": ["<surface-id>"], "adversarialCaseRefs": ["<case-id>"] }
    ],
    "surfaceEvidence": [
      { "id": "<surface-id>", "contractRef": "<surface under test>", "surface": "gui|web|cli|api|package|algorithm|math|native|desktop|tui", "invocation": "<real invocation>", "verdict": "passed", "artifactRefs": ["<ref-id>"] }
    ],
    "adversarialCases": [
      { "id": "<case-id>", "contractRef": "<approved contract id>", "scenario": "<boundary/adversarial input>", "expectedBehavior": "<required handling>", "verdict": "passed", "artifactRefs": ["<ref-id>"] }
    ],
    "blockers": []
  },
  "iteration": {
    "status": "passed",
    "evidence": "blockers absent or resolved and the full loop was rerun cleanly",
    "fullRerun": true,
    "rerunCommands": ["bun test:e2e", "bun test:red-team"],
    "reviewCohort": {
      "reviewGeneration": 1,
      "sourceHash": "sha256:<frozen change-set hash every lane inspected>",
      "joined": true,
      "lanes": {
        "cleaner": { "status": "passed", "sourceHash": "sha256:<same>", "evidence": "AI SLOP CLEANUP REPORT: zero blocking findings", "blockers": [] },
        "architect": { "status": "CLEAR", "sourceHash": "sha256:<same>", "evidence": "architecture/product/code review of the frozen set", "blockers": [] },
        "qa": { "status": "passed", "sourceHash": "sha256:<same>", "evidence": "e2e + red-team run against the frozen set", "blockers": [] }
      }
    },
    "blockers": []
  }
}

Provide one artifactRefs entry per live surface actually exercised, using the surface-appropriate kind and evidence rules from steps 6–7 above; the CLI rejects missing or shallow gates. status: "not_applicable" rows are allowed only in contractCoverage and surfaceEvidence and each requires contractRef plus reason.

For safe CLI replay artifacts, the JSON at path must be an object like {"schemaVersion":1,"kind":"cli-replay","replaySafe":true,"command":["bun","-e","console.log(\"ultragoal-cli-ok\")"],"cwd":".","env":{"LC_ALL":"C"},"timeoutMs":30000,"expectedExitCode":0,"recordedStdout":"ultragoal-cli-ok\n","recordedStderr":"","invariants":[{"type":"substring","value":"ultragoal-cli-ok"},{"type":"not_substring","value":"error"}]}. replaySafe: true is required but is never authority by itself: executable replay is limited to the pinned Bun runtime for bun --version or deterministic literal bun -e "console.log(...)". Shells, nested interpreters, path-qualified executables, test source, install/publish commands, git mutation, network clients, and every other argv are rejected. The declared cwd and artifact files are realpath-confined beneath the repository, but the safe probe itself runs from a fresh empty temporary cwd/home so repository bunfig.toml preloads and user configuration cannot execute. Mixed inline/nested/file-backed rows fail closed, POSIX timeout cleanup signals the replay process group, stdout and stderr are validated after normalization, output is capped at 1 MiB, and the child environment is scrubbed to CI, NO_COLOR, GJC_ULTRAGOAL_REPLAY, trusted temporary HOME/TMPDIR, plus optional LANG, LC_ALL, LC_CTYPE, and TZ.

Compiled GJC binaries fail executable replay closed because their process.execPath launches GJC rather than a Bun CLI. Those runs must use the existing audited replayExempt structural fallback; the validator never resolves an untrusted bun from PATH.

Focused bun test execution is blocked because repository test source is still arbitrary host code without an operating-system sandbox. The current replayExempt contract continues to require an existing screenshot, automation transcript, or PTY structural fallback; a test-report or bun-test-report JSON file is intentionally not accepted yet. Keep that design work open until test-result provenance, output binding, and consumer authority can be made fail-closed. Allowed reasonCode values remain unsafe_side_effect, requires_credentials, requires_network, non_deterministic_external, destructive, interactive_only, and platform_unavailable.

Terminal critic gate

The terminal critic gate is a fail-closed, once-per-run-terminus review. It guards both terminal exits with a read-only critic role agent's OKAY verdict; it does not run per story. It is additive to, and does not change, the existing per-story architect review and executor QA/red-team lanes.

Completion terminus

Before assembling the final-aggregate --quality-gate-json, the leader delegates the terminal critic. Only the final-aggregate completion checkpoint requires the additional top-level criticReview key; criticReview is tolerated but ignored on non-final checkpoints. A clean final aggregate requires verdict: "OKAY", non-empty evidence, and an empty blockers array:

{
  "criticReview": {
    "verdict": "OKAY",
    "evidence": "terminal critic review of the final required-goal state",
    "blockers": []
  }
}

Pause/blocked terminus

At a human_blocked terminus, the leader first runs gjc ultragoal classify-blocker --classification human_blocked (capturing that classification's ledger eventId), then delegates the terminal critic and records its verdict with gjc ultragoal record-critic-verdict --terminus pause --classification-event-id <eventId> before calling goal({"op":"pause"}). The pause is allowed only when a later fresh critic_verdict ledger receipt exists with terminus: "pause", verdict: "OKAY", non-empty evidence, an empty blockers array, the current planGeneration, and a classificationEventId bound to the latest blocker_classified event, which must be human_blocked. Freshness is scoped to the final required-goal state, so required-goal or steer changes stale the receipt, and a newer classification supersedes an older verdict.

The critic must verify that the human_blocked classification is genuine, including catching false pauses where needed resources exist locally or the asserted blocker is resolvable. A REJECT (or ITERATE) verdict refuses the terminal pause; the run keeps executing. The pause (goal({"op":"pause"})) is the gated terminal park-and-wait exit — a per-goal gjc ultragoal checkpoint --status blocked remains available as non-terminal blocker bookkeeping that never signals run completion and keeps the blocker outstanding until resolved.

Invocation and containment

At each terminus, the leader gives the read-only critic role agent brief.md, goals.json, ledger.jsonl, and the cumulative change set. For completion, invoke it before assembling the final-aggregate gate JSON. For pause, invoke it after the human_blocked classification and before goal({"op":"pause"}). The terminal critic must not spawn nested ralplan, deep-interview, or ultragoal workflows. This creates no interactive surface: ask remains blocked while an Ultragoal run is active.

On repeat terminus attempts within the same run (after an ITERATE/REJECT reopen cycle or a superseded pause classification), resume the prior terminal-critic subagent when resumable instead of freshly spawning one: the critic already holds brief.md, goals.json, the ledger history, and its own prior findings, so re-invocation only needs the delta (new ledger events, the updated cumulative change set, and evidence addressing the prior blockers). Resume via existing subagent resume/steer controls; on context_unavailable, not_found, no_runner, or resume_failed — or after a process restart — fall back to a fresh critic spawn with the full context bundle. A resumed terminal critic remains read-only, keeps the same containment rules, and must issue a fresh verdict against the current state — a prior ITERATE is never carried forward as pre-judged, and each verdict is still recorded through gjc ultragoal record-critic-verdict.

Non-OKAY loop and ceiling

For completion-side ITERATE or REJECT, the leader MUST first record the terminal verdict so the run-level counter observes it: gjc ultragoal record-critic-verdict --terminus completion --verdict <ITERATE|REJECT> --evidence "<critic findings>"; then record the findings with gjc ultragoal record-review-blockers and reopen the run. The dedicated counter ceiling is 5, independently of the give-up nudge budget, and is RUN-LEVEL: it counts every non-OKAY terminal-critic verdict across the whole run and all reopen cycles. On reaching that ceiling, both pause and final completion are blocked until a human or leader records gjc ultragoal record-critic-gate-override --evidence "<authorization evidence>". There is no automatic pause override.

This gate is always fail-closed and has no grandfathering: in-flight runs must obtain a terminal verdict when they reach a terminus.

Deferred / out of scope

Gating active-aggregate goal drop after nudge exhaustion is a known follow-up not covered by this gate; drop remains governed by the existing nudge discipline.

Review mode

gjc ultragoal review runs the same hardened gate against an already implemented PR, branch, or worktree. Use --pr <number> for a PR, --branch <ref> for a branch diff, omit both for the current worktree, and pass --spec <path> when a real contract exists. --mode review-only emits the verdict/findings without creating fix work; --mode review-start records review blockers for follow-up. Review mode validates the same executorQa shape and live-surface artifacts as checkpoint --status complete. A thin or derived-only contract can never clean-pass: the verdict is capped at inconclusive: weak-contract until a supplied spec or equivalent strong acceptance criteria are available.

Receipts are freshness-scoped:

• Per-goal receipts remain fresh for their target goal unless that goal, its blocker metadata, or its supersession metadata changes.
• Normal later goal_started or clean receipt-backed goal_checkpointed events for other goals do not stale older per-goal receipts.
• Appending required goals or changing final required-goal state stales final aggregate receipts. Final aggregate completion requires a fresh final aggregate receipt proving no incomplete, blocked, or review_blocked required goals remain.
• Deferred per-goal receipts (validation-batch members) are incomplete until a matching fresh batch-close receipt exists on the batch's finalGoalId; a story-scope query for a deferred member stays blocked until that close, and mutating a member after close stales the batch-close and final aggregate receipts.

Cross-repository succession

When an approved run was planned in one repository but the implementation belongs in another, the repository-binding guard correctly refuses to execute the plan from the second repository. Never edit repositoryBinding and never copy the runtime directory: both defeat the boundary and launder verification provenance. Use the explicit successor path instead.

gjc ultragoal succession offer --target-repo <path> --goal-id G002 --goal-id G003 --authorize "<who approved what, for which source, target and goals>" --authorized-by <identity> --json
gjc ultragoal succession adopt --offer <offer path inside the source worktree> --json
gjc ultragoal succession status --json

offer runs in the source repository. It records an offer document plus a durable outgoing ownership fence, and writes nothing to brief.md, goals.json or ledger.jsonl. From that moment the source run refuses to schedule or checkpoint the fenced goals; goals outside the selection stay fully schedulable. Ownership therefore leaves the source at the offer, not when the target happens to adopt, so no interval exists in which both runs may resume the same goal. There is deliberately no revoke verb: any read-then-delete of the fence races a concurrent adoption and could revive the source while the target is starting, so an unadopted offer keeps its goals fenced.

Because unselected goals stay schedulable, ordinary source progress moves whole-file digests. Adoption therefore compares a snapshot of the offered work — the brief plus each selected goal's record and unresolved obligations — instead of whole-file equality, so unrelated progress cannot permanently strand the handed-off work. Whole-file digests are still recorded as provenance and any drift is written into the adoption record.

adopt runs in the target repository and starts a genuinely fresh run: new pending goals, fresh timestamps, the target repository's own binding, and no inherited receipts, quality gates, validation batches, steering or activated objective. The source brief and every selected objective are carried verbatim, together with the unresolved obligations recorded against those goals — requirements are carried forward in full, never summarized down. Dependency groups travel as explicit obligations with source-to-target id mapping: the source's batch metadata and receipts are not inherited, but the requirement to validate those goals together is.

Selection and authorization must be explicit. Only unfinished goals (pending, failed, blocked, review_blocked) may be handed off; the source run must be quiescent with no active goal; there is no implicit or wildcard selection; and a selection that takes only part of a validation batch, or splits a review blocker from the goal it still blocks, is rejected rather than silently dropping that obligation. Adoption fails closed on changed offered work, an altered or unfenced offer, an offer path that escapes the source worktree, a target the offer does not name, a different linked worktree from the exact named target, an occupied target session, and any duplicate or simultaneous adoption anywhere in the target repository. A linked worktree is a valid target when the offer names that exact worktree; the source and target may therefore be sibling worktrees, but aliases of the same source worktree remain self-transfers and are rejected using physical path identity, and an unavailable bound worktree root is refused rather than compared lexically. A retry reconciles the exact recorded operation or fails closed — a matching operation id is never authority to overwrite target work it did not write, and a published plan that has disappeared is never reconstructed.

Until publication completes, the target refuses to execute: a visible goals.json is not admission evidence, the adoption record is. start/complete-goals and checkpoint reject while an adoption remains pending.

Integrity digests are evidence, not authorization: a matching digest set with no fence, no bounded authorization, or the wrong target is still rejected. Historical source statuses, receipts, reviews and approvals travel as provenance only and never satisfy a target gate.

Handoff back to planning

When the aggregate ultragoal is complete OR the user requests return to planning/clarification, mark ultragoal ready for handoff so the skill tool's chain guard permits the backward transition:

gjc state ultragoal write --input '{"current_phase":"handoff"}' --json

The skill tool then dispatches /skill:ralplan or /skill:deep-interview same-turn and runs gjc state ultragoal handoff --to <ralplan|deep-interview> --json in-process to atomically demote ultragoal, promote the callee, and sync both .gjc/_session-{sessionid}/state/skill-active-state.json files. You do not need to run the handoff verb yourself.

Constraints

• The shell command emits a model-facing handoff for the active GJC agent; it does not invoke any /goal slash-command and the agent loop must not depend on any /goal subcommand.
• Use only the unified goal-tool surface from the agent loop: goal({"op":"get"}), goal({"op":"create"}), goal({"op":"complete"}), goal({"op":"drop"}), goal({"op":"resume"}). drop clears the active goal without exiting goal mode so the next goal({"op":"create"}) works in-session. No slash-command cleanup exists or is required; Ultragoal never calls any /goal subcommand.
• For back-to-back ultragoal runs in the same session/thread, when goal({"op":"get"}) still reports an active aggregate, call goal({"op":"drop"}) before goal({"op":"create"}); when no active goal exists or the prior aggregate is already complete or dropped, call goal({"op":"create"}) directly. The goal tool remains callable across drop; no slash-command cleanup exists or is required.
• Never call goal({"op":"create"}) when goal({"op":"get"}) reports a different active goal.
• Never call goal({"op":"complete"}) unless the aggregate run or legacy per-story goal is actually complete.
• In aggregate mode, intermediate and final story checkpoints update durable goals.json state and append receipt proof to ledger.jsonl; the final story checkpoint creates the final aggregate receipt before the agent may call goal({"op":"complete"}).
• Completion checkpoints require --quality-gate-json only. Shell commands and hooks must not mutate goal state; the agent reconciles inline goal-tool state after durable completion.
• Final-aggregate completion additionally requires a criticReview OKAY; a human_blocked pause additionally requires a fresh OKAY critic_verdict receipt.
• Treat ledger.jsonl as the durable audit trail; checkpoint after every success or failure.