KRT Swarm Seneschal
Coordinate delivery only when several work units benefit from shared planning,
isolation, parallel execution, or cross-unit reconciliation:
source work -> plan -> dispatch -> reconcile -> release handoff
This skill coordinates other workflows. It does not replace
krt-compound-master, Jira provider skills, or krt-release-marshal.
Entry Gate
Apply the break-even gate before creating swarm state or loading detailed
references:
- Keep one small, execution-ready unit in the root task.
- Route generic planning that does not require worker-wave decomposition, plus
Jira-only, review-only, and release-only requests, to the skill that owns the
requested outcome.
- Use Seneschal only for multiple dependent or concurrent units, multiple
isolated Compound Master runs, or an explicitly requested worker wave.
Before any code, repository-state, queue-state, Jira, branch, PR, or release
mutation, read references/safety.md.
Four Operations
| Operation |
Outcome |
plan |
Decide root-direct versus swarm, normalize units, close dependencies, and select the smallest safe wave. |
dispatch |
Launch only the admitted workers with bounded ownership, verification, and isolation. |
reconcile |
Inspect real changes and evidence, resolve findings or blockers, and prepare release-ready handoff. |
status |
Render current gates, units, blockers, and next action without mutation. |
Legacy names remain accepted as aliases, not separate pipelines:
design-only, document-plan, document-review, document-revise,
document-approve, wave-plan, and jira-seed-and-drain start in plan.jira-team-flow, overnight-team-flow, and autonomous-team-flow apply an
external-source or autonomy overlay to dispatch.resume begins with status, then continues through plan, dispatch, or
reconcile from observed state.blocker-review is status; blocker-resolve is reconcile.wave-status is status.
Operating Model
Plan
- Inspect the repository, source work, live worker facts, and existing state.
Before creating a wave, load
references/worktree-collaboration.md and run
its guarded preflight cleanup in dry-run mode. Apply cleanup only to entries
already marked cleanup-ready by canonical state; retain and report every
active, failed, diagnostic, or unregistered workspace.
- Reject work that lacks checkable acceptance criteria, safe ownership, known
dependencies, or a resolvable decision boundary.
- Require a reviewed documentation packet before seeding Jira or dispatching
a broad initiative, roadmap, program, or rough backlog. Do not manufacture a
new packet for an already approved, execution-ready unit explicitly requested
by the user; when persisted, bind that exemption to the trusted request as
specified in
references/queue-state-schema.md.
- Load
references/execution-lanes.md; classify execution difficulty and
assurance independently.
- Prefer the smallest useful wave. Default to at most two concurrent mutable
implementers and serialize overlapping auth, data, migration, public-contract,
central-model, dependency, or lockfile work.
Dispatch
Every worker receives a bounded objective, owned paths, non-goals, acceptance
criteria, exact focused checks, stop conditions, and a finite time/action budget.
Workers do not stage, commit, switch branches, manage worktrees, push, mutate
Jira, request reviewers, open PRs, or merge.
Choose the least costly protocol that preserves the required assurance:
| Condition |
Protocol |
low assurance |
Root-direct by default; use lightweight dispatch only for a named isolation, concurrency, or duration benefit. |
medium assurance |
Lightweight dispatch plus one focused independent reviewer. |
high assurance, deep execution, or autonomous mutation |
Executable hashed contract, root observation, focused specialist review, and independent validation. |
critical assurance |
High protocol plus coordinated review and explicit approval or an exact ledger grant. |
Use references/lightweight-dispatch.md for eligible low/medium work. Load
references/executable-worker-contracts.md and references/subagent-contracts.md
only for the advanced protocol. A worker's prose never certifies its real diff
or readiness; root owns that observation in both protocols.
Reconcile
- Inspect the actual diff and changed paths before trusting a worker return.
- Capture or rerun the checks required by the selected protocol. Run aggregate
or CI-equivalent verification once for the reconciled wave; reuse only
content-bound passing evidence.
- Apply only risk-triggered review. Use the findings registry and coordinated
validators only for high or critical assurance.
- Mark each unit
release-ready, needs-fix, blocked, deferred, or
split-required. Continue independent work when a blocker is local.
- Once an invocation's patch, manifest, and required evidence are durable,
mark its workspace cleanup-ready and immediately run the guarded cleanup.
Preserve failed or contract-violating workspaces for diagnosis.
- Hand release-ready work to
krt-release-marshal; do not reproduce its
commit, PR, Jira-transition, reviewer, or merge workflow.
Status
Read canonical local state and live external facts when relevant. Run
scripts/render_swarm_status.py for the compact panel. Do not create or mutate
state merely to report status.
Progressive Reference Router
Read the primary reference for the active operation first. Load a secondary
reference only when its trigger is present.
| Trigger |
Reference |
| Plan or select a wave |
references/queue-and-dispatch.md |
| Rough initiative or documentation approval |
references/documentary-planning.md |
| Nested Compound Master runs |
references/compound-master-nesting.md |
| Execution lane, assurance, or role admission |
references/execution-lanes.md |
| Eligible low/medium worker |
references/lightweight-dispatch.md |
| High/critical, deep, autonomous, or policy-required worker |
references/executable-worker-contracts.md |
| Complex role prompts or named worker profiles |
references/subagent-contracts.md, then references/worker-profiles.md if needed |
| Concurrent mutable workers |
references/worktree-collaboration.md |
| Coupled work around a small shared foundation |
references/staged-decomposition.md |
| Adaptive concurrency or reusable aggregate evidence |
references/automated-wave-control.md |
| Reconciliation and release readiness |
references/gates-and-reconciliation.md |
| High/critical coordinated review |
references/review-coordination.md |
| Interrupted review or disputed verification |
references/role-recoverability.md |
| Persistent queue or blockers |
references/queue-state-schema.md, then references/blocker-ledger.md if needed |
| Jira source or seeding |
references/jira-team-flow.md, then references/jira-seeding.md if needed |
| No-confirmation run |
references/autonomous-team-flow.md |
| Design explanation only |
references/swarm-protocol.md |
Stop Conditions
Stop the affected path when authority is missing, a required documentation gate
is not approved, ownership overlaps unsafely, a high-risk decision is unresolved,
verification fails, or the selected Jira provider is ambiguous. In autonomous
flow, record the blocker and continue only independent ledger-covered work.
Closeout
Return the operation, gate status when applicable, units and outcomes, root-owned
verification/review evidence, blockers, branch/worktree references, changed
state artifacts, and one exact next action. Omit machinery that the selected
assurance protocol did not use.
1---2name: krt-swarm-seneschal3description: Coordinate explicitly requested multi-worker delivery through dependency-aware planning, isolated worker waves, reconciliation, and release handoff. Use for swarm orchestration, parallel Codex workers, or multiple nested krt-compound-master runs. Do not use for one execution-ready task, generic project planning, Jira administration, or release management.4---56# KRT Swarm Seneschal78Coordinate delivery only when several work units benefit from shared planning,9isolation, parallel execution, or cross-unit reconciliation:1011```text12source work -> plan -> dispatch -> reconcile -> release handoff13```1415This skill coordinates other workflows. It does not replace16`krt-compound-master`, Jira provider skills, or `krt-release-marshal`.1718## Entry Gate1920Apply the break-even gate before creating swarm state or loading detailed21references:22231. Keep one small, execution-ready unit in the root task.242. Route generic planning that does not require worker-wave decomposition, plus25 Jira-only, review-only, and release-only requests, to the skill that owns the26 requested outcome.273. Use Seneschal only for multiple dependent or concurrent units, multiple28 isolated Compound Master runs, or an explicitly requested worker wave.2930Before any code, repository-state, queue-state, Jira, branch, PR, or release31mutation, read [references/safety.md](references/safety.md).3233## Four Operations3435| Operation | Outcome |36|---|---|37| `plan` | Decide root-direct versus swarm, normalize units, close dependencies, and select the smallest safe wave. |38| `dispatch` | Launch only the admitted workers with bounded ownership, verification, and isolation. |39| `reconcile` | Inspect real changes and evidence, resolve findings or blockers, and prepare release-ready handoff. |40| `status` | Render current gates, units, blockers, and next action without mutation. |4142Legacy names remain accepted as aliases, not separate pipelines:4344- `design-only`, `document-plan`, `document-review`, `document-revise`,45 `document-approve`, `wave-plan`, and `jira-seed-and-drain` start in `plan`.46- `jira-team-flow`, `overnight-team-flow`, and `autonomous-team-flow` apply an47 external-source or autonomy overlay to `dispatch`.48- `resume` begins with `status`, then continues through `plan`, `dispatch`, or49 `reconcile` from observed state.50- `blocker-review` is `status`; `blocker-resolve` is `reconcile`.51- `wave-status` is `status`.5253## Operating Model5455### Plan56571. Inspect the repository, source work, live worker facts, and existing state.58 Before creating a wave, load `references/worktree-collaboration.md` and run59 its guarded preflight cleanup in dry-run mode. Apply cleanup only to entries60 already marked cleanup-ready by canonical state; retain and report every61 active, failed, diagnostic, or unregistered workspace.622. Reject work that lacks checkable acceptance criteria, safe ownership, known63 dependencies, or a resolvable decision boundary.643. Require a reviewed documentation packet before seeding Jira or dispatching65 a broad initiative, roadmap, program, or rough backlog. Do not manufacture a66 new packet for an already approved, execution-ready unit explicitly requested67 by the user; when persisted, bind that exemption to the trusted request as68 specified in `references/queue-state-schema.md`.694. Load `references/execution-lanes.md`; classify execution difficulty and70 assurance independently.715. Prefer the smallest useful wave. Default to at most two concurrent mutable72 implementers and serialize overlapping auth, data, migration, public-contract,73 central-model, dependency, or lockfile work.7475### Dispatch7677Every worker receives a bounded objective, owned paths, non-goals, acceptance78criteria, exact focused checks, stop conditions, and a finite time/action budget.79Workers do not stage, commit, switch branches, manage worktrees, push, mutate80Jira, request reviewers, open PRs, or merge.8182Choose the least costly protocol that preserves the required assurance:8384| Condition | Protocol |85|---|---|86| `low` assurance | Root-direct by default; use lightweight dispatch only for a named isolation, concurrency, or duration benefit. |87| `medium` assurance | Lightweight dispatch plus one focused independent reviewer. |88| `high` assurance, `deep` execution, or autonomous mutation | Executable hashed contract, root observation, focused specialist review, and independent validation. |89| `critical` assurance | High protocol plus coordinated review and explicit approval or an exact ledger grant. |9091Use `references/lightweight-dispatch.md` for eligible low/medium work. Load92`references/executable-worker-contracts.md` and `references/subagent-contracts.md`93only for the advanced protocol. A worker's prose never certifies its real diff94or readiness; root owns that observation in both protocols.9596### Reconcile97981. Inspect the actual diff and changed paths before trusting a worker return.992. Capture or rerun the checks required by the selected protocol. Run aggregate100 or CI-equivalent verification once for the reconciled wave; reuse only101 content-bound passing evidence.1023. Apply only risk-triggered review. Use the findings registry and coordinated103 validators only for high or critical assurance.1044. Mark each unit `release-ready`, `needs-fix`, `blocked`, `deferred`, or105 `split-required`. Continue independent work when a blocker is local.1065. Once an invocation's patch, manifest, and required evidence are durable,107 mark its workspace cleanup-ready and immediately run the guarded cleanup.108 Preserve failed or contract-violating workspaces for diagnosis.1096. Hand release-ready work to `krt-release-marshal`; do not reproduce its110 commit, PR, Jira-transition, reviewer, or merge workflow.111112### Status113114Read canonical local state and live external facts when relevant. Run115`scripts/render_swarm_status.py` for the compact panel. Do not create or mutate116state merely to report status.117118## Progressive Reference Router119120Read the primary reference for the active operation first. Load a secondary121reference only when its trigger is present.122123| Trigger | Reference |124|---|---|125| Plan or select a wave | `references/queue-and-dispatch.md` |126| Rough initiative or documentation approval | `references/documentary-planning.md` |127| Nested Compound Master runs | `references/compound-master-nesting.md` |128| Execution lane, assurance, or role admission | `references/execution-lanes.md` |129| Eligible low/medium worker | `references/lightweight-dispatch.md` |130| High/critical, deep, autonomous, or policy-required worker | `references/executable-worker-contracts.md` |131| Complex role prompts or named worker profiles | `references/subagent-contracts.md`, then `references/worker-profiles.md` if needed |132| Concurrent mutable workers | `references/worktree-collaboration.md` |133| Coupled work around a small shared foundation | `references/staged-decomposition.md` |134| Adaptive concurrency or reusable aggregate evidence | `references/automated-wave-control.md` |135| Reconciliation and release readiness | `references/gates-and-reconciliation.md` |136| High/critical coordinated review | `references/review-coordination.md` |137| Interrupted review or disputed verification | `references/role-recoverability.md` |138| Persistent queue or blockers | `references/queue-state-schema.md`, then `references/blocker-ledger.md` if needed |139| Jira source or seeding | `references/jira-team-flow.md`, then `references/jira-seeding.md` if needed |140| No-confirmation run | `references/autonomous-team-flow.md` |141| Design explanation only | `references/swarm-protocol.md` |142143## Stop Conditions144145Stop the affected path when authority is missing, a required documentation gate146is not approved, ownership overlaps unsafely, a high-risk decision is unresolved,147verification fails, or the selected Jira provider is ambiguous. In autonomous148flow, record the blocker and continue only independent ledger-covered work.149150## Closeout151152Return the operation, gate status when applicable, units and outcomes, root-owned153verification/review evidence, blockers, branch/worktree references, changed154state artifacts, and one exact next action. Omit machinery that the selected155assurance protocol did not use.
elzawarudo