@alphazede/bearing-lite 0.1.0 → 0.1.1

package/skills/map-the-route/SKILL.md

@@ -1,33 +1,43 @@
 ---
 name: map-the-route
 description: >
-  Build the human-readable task graph when the outcome is approved but
-  executable tasks are missing. Use for map the route, implementation slices,
-  task graph, SEIT, or planning artifacts. Do not use to execute work, select
-  models, self-certify assurance, or advance journey state beyond planning.
+  Create or resume Bearing's topic-specific specification, design, SEIT,
+  implementation graph, and self-contained review HTML after decisions are
+  resolved. Use for Map the Route or missing planning artifacts. Do not use to
+  implement, select agents/models, self-approve, or bypass owner review.
 ---
 
 # Map the Route
 
-Procedural stage. Not a persona.
-
-## Match / non-match
-
-- **Match:** plan-spec/decisions are approved and executable task graph or planning package is incomplete.
-- **Non-match:** tasks already READY with roles and assurance; pure Crewmate packet work.
-
-## Inputs
-
-Approved goal, decisions, plan-spec, and required Role routes.
-
-## Procedure
-
-1. Author only missing planning artifacts: design, SEIT, and implementation slices as needed.
-2. For each task declare `depends_on`, assigned role, scope, authority, and `required_assurance` (explicit `none` allowed).
-3. Prefer the smallest valid route; leave dormant roles unselected.
-4. Keep IDs stable; cite acceptance and evidence expectations per task.
-5. Return artifact paths, blockers, and next planning or execution handoff without executing.
-
-## Never
-
-Implement product changes, invent owner approval, run phase gates as automatic, or claim independent assurance.
+Fresh planning node. The Router remains the plan-state writer and user contact.
+
+## Inputs and match
+
+- **Inputs:** confirmed decisions, repository map, artifact status, requirements,
+  owner-selected lineup/cadence, repository rules, and return schema.
+- **Match:** first specification gate or post-approval planning package is missing.
+- **Non-match:** executable tasks already map to current approved artifacts.
+
+## Algorithm
+
+1. Derive `<journey-topic>-spec.md` from the confirmed title and return the name
+   for owner verification before first write. Never require generic `plan-spec.md`.
+2. Write the testable specification and self-contained offline `review.html`.
+   Open the exact HTML in the user's browser and verify the launcher/process.
+3. After specification approval, select and record suitable design lenses
+   without asking unless owner intent changes. Write `design.md`, then `seit.md`;
+   draft `implementation.md` only after both prospective gates are complete.
+4. Give every slice stable requirement/design/SEIT IDs, dependencies, exact
+   write set, authority, fresh-session role, evidence, recovery, and stop rule.
+5. Copy the confirmed lineup and review cadence into implementation and HTML.
+   Show named active, standby, and unused role instances.
+6. Generate the final HTML with deep zoomable diagrams for architecture, state,
+   interactions, failures, dependencies, and ownership. Put the recommended
+   Journey path in Implementation, open it, and request owner approval.
+
+## Return and recovery
+
+Return `SPEC_REVIEW_READY`, `PLAN_REVIEW_READY`, `NEEDS_OWNER_DECISION`, or
+`VALIDATION_FAILED` with artifact paths, evidence, blocker, and next action.
+After final approval, `review.html` is authoritative. At most three
+evidence-changing correction rounds; never invent approval or implement.

package/skills/map-the-route/references/artifact-grammar.md

@@ -10,13 +10,13 @@ Bearing rejects deviations with typed findings.
 2. Requirement IDs use `AC-*` or `RISK-*`; design IDs use `DES-*` or
    `CONTRACT-*`; SEIT rows use `SEIT-*`; commands use `CMD-*` or `PROC-*`.
    Suffixes contain only uppercase letters, digits, dots, or hyphens.
-3. `plan-spec.md` declares requirements, Entry criteria, Exit criteria,
+3. `<journey-topic>-spec.md` declares requirements, Entry criteria, Exit criteria,
    Rollback or repair, and Accountable controller. `design.md` declares design
    IDs. `seit.md` declares SEIT rows and commands.
 4. Preserve stable IDs. Every implementation reference resolves to a declared
    requirement, design contract, SEIT row, and command.
-5. The plan's Role routes decision supplies immutable `authorRoute` and
-   `reviewSlots`. Missing or incomplete routes block implementation drafting.
+5. The owner-confirmed lineup and review cadence supply named active, standby,
+   and unused instances. Missing values block implementation drafting.
 
 ## SEIT rules
 
@@ -38,8 +38,8 @@ Bearing rejects deviations with typed findings.
    `### <id> execution manifest`. IDs are a letter-run plus an integer or
    dotted integer. Every slice has exactly one matching manifest.
 2. Every slice declares Goal, Requirement IDs, Design IDs, SEIT proof rows,
-   Type, Design lenses, Implementation role, Agent model route, Agent reasoning
-   level, and Review path. Goals are at most 512 characters.
+   Type, Design lenses, named implementation role, owner-selected model route,
+   reasoning level, and review path. Goals are at most 512 characters.
 3. Every manifest declares Write set, Command IDs, Stop condition, and Human
    decision. Optional fields are Shared interfaces (`path#Symbol`), Integration
    boundary, and Parallel safe (`yes` or `no` plus reason).
@@ -76,8 +76,8 @@ placeholders, bare negations, or deferral language.
 
 ## Completion boundary
 
-Design-only work returns after valid `design.md` and `seit.md`. Implementation
-drafting returns after valid `implementation.md`. Complete-route work remains
-incomplete until `plan-spec.md`, `design.md`, `seit.md`, and `implementation.md`
-are current. Bearing then generates the complete `review.html`; agents never
-hand-edit, summarize, or replace it.
+The first review gate needs the topic-specific specification and `review.html`.
+After approval, complete-route work remains incomplete until that specification,
+`design.md`, `seit.md`, and `implementation.md` are current. Bearing then
+generates and opens the complete `review.html`; it becomes authoritative only
+after owner approval.

package/skills/navigator/SKILL.md

@@ -1,44 +1,42 @@
 ---
 name: navigator
 description: >
-  Expedition and multi-phase Navigator sequencing across dependency-connected
-  waves, including gap correction and in-authority coverage expansion. Use for
-  navigator, expedition integration boundary, or multi-phase wave graph control.
-  Do not use for single bounded packets, product implementation under Navigator
-  identity, or independent assurance of own corrections.
+  Sequence dependency-connected Expedition waves, fresh-session coordinators,
+  and integration boundaries. Use for Navigator, Expedition, or multi-wave
+  execution control. Do not use for one packet/wave, product implementation,
+  planning-state ownership, model selection, publication, or assurance.
 ---
 
 # Navigator
 
-## Trigger
+Highest execution authority below Router; performs the least hands-on execution.
 
-- **Match:** work spans multiple phases, dependency-connected waves, or an Expedition integration boundary.
-- **Non-match:** one packet or one Explorer-owned wave with no cross-wave sequencing need.
+## Inputs and match
 
-## Inputs
+- **Inputs:** approved artifacts, wave graph, dependencies, repository scope,
+  authority, lineup, review cadence, acceptance, and return schema.
+- **Match:** an Expedition has multiple dependency-connected waves or phases.
+- **Non-match:** one Explorer wave or Crewmate packet is sufficient.
 
-Plan path, wave graph, dependencies, approved outcome, authority, allowed paths, `required_assurance`, expected handoff.
+## Algorithm
 
-## Responsibility
+1. Start fresh and reread the approved specification, design, SEIT,
+   implementation, and review baseline. Verify snapshot lineup and cadence.
+2. Sequence fresh Explorer sessions. Add Trail Boss only for simultaneous waves
+   with shared-interface or dependency conflicts.
+3. Integrate typed wave returns and correct only sequencing/coverage gaps inside
+   approved authority; never perform implementation.
+4. Dispatch independent assurance at the confirmed cadence: every slice, every
+   integrated execution/correction round, or only the final integrated outcome.
+5. Return to the Router after the bounded Expedition outcome; do not silently
+   continue into another Journey or protected action.
 
-Sequence Explorers (and Trail Boss only when multi-wave conflict requires it), correct in-authority gaps, and integrate returns. Use one fresh Navigator context per execution wave; after phase acceptance PASS, return to Delegate Authority rather than continuing into the next wave. A fresh wave Navigator rereads `implementation.md`, `review.html`, and `seit.md` before acting. Never implement under Navigator identity, supply independent assurance, publish, release, mutate remotes, or take destructive action under Navigator identity.
+## Return and recovery
 
-## Authority
+Return `READY`, `REROUTED`, `WAITING_ON`, or `OWNER_DECISION_REQUIRED` with
+plan ref, role, subject, dependencies, scope, authority, evidence, blocker, next
+action, and receiver. Attempts 1–3 require new evidence or strategy; a material
+design, intent, scope, security, destructive, remote, or publication change
+returns to Owner Authority.
 
-Approved plan plus task authority is sufficient for execution, routing, bounded repairs, and in-contract coverage amendments. Do not re-request Owner Authority for work already authorized by that plan.
-
-Before any `OWNER_DECISION_REQUIRED` return, seek and attempt a safer in-authority workaround when one exists. Ask Owner Authority only when outcome or acceptance meaning changes; repository or product boundary changes; public, publication, remote, destructive, release, or release-target action is needed; secret handling or a security exception changes; an external dependency or cost requires new authority; a confirmed owner decision would change; any other authority envelope would expand; or the declared third failed correction occurs.
-
-## Return
-
-Handoff fields: `plan_ref`, `role`, `subject`, `depends_on`, `scope`, `authority`, `outcome`, `evidence`, `blocker`, `next_action`, `receiving_role`.
-
-Outcomes: `READY`, `REROUTED`, `WAITING_ON`, `OWNER_DECISION_REQUIRED`. Receiver: Trail Boss, Explorer, assurance role when required, Delegate Authority, or Owner Authority.
-
-## Independence
-
-Writes only coordinator sequencing and task blocks it owns. Does not supply Validator, Park Ranger, or Surveyor verdicts for candidates it integrated as author-equivalent.
-
-## Correction
-
-Attempts 1–2 need new hypothesis and evidence. Third failure or out-of-authority change → `OWNER_DECISION_REQUIRED`.
+Never implement, self-assure, select models, or mutate remotes.

package/skills/park-ranger/SKILL.md

@@ -1,38 +1,39 @@
 ---
 name: park-ranger
 description: >
-  Review introduced defects on one exact candidate when required_assurance
-  includes Park Ranger, the owner opts in, or a material integrated phase gate
-  requires defect review. Use for park ranger, defect review, or introduced-bug
-  adjudication. Do not use for implementation, automatic per-slice gates,
-  sufficiency scoring, or user-facing acceptance.
+  Independently review one exact stable candidate for introduced actionable
+  defects in a fresh session at the selected review boundary. Use for Park
+  Ranger, code review, or defect adjudication. Do not use for implementation,
+  automatic review, evidence scoring, user acceptance, or author self-review.
 ---
 
 # Park Ranger
 
-## Trigger
+Independent defect assurance, outside the mutation-authority ladder.
 
-- **Match:** `required_assurance` lists Park Ranger; owner opts in; or integrated phase gate requires defect review on a stable candidate.
-- **Non-match:** `required_assurance: none` without owner opt-in/phase gate; Validator sufficiency-only; Surveyor acceptance; implementation packets.
+## Inputs and match
 
-## Inputs
+- **Inputs:** approved baseline, exact candidate ref and diff, author identity,
+  relevant evidence, cadence boundary, review focus, and return schema.
+- **Match:** Park Ranger is declared and the selected `per-slice`, `per-round`,
+  or `at-end` boundary has a stable candidate.
+- **Non-match:** boundary is not reached, candidate is unstable/unchanged, or
+  Validator/Surveyor work is requested.
 
-Plan path, exact `candidate_ref`, diff or evidence, prior Validator result when present, expected handoff.
+## Algorithm
 
-## Responsibility
+1. Start fresh; reject author identity, candidate discontinuity, or missing
+   review boundary.
+2. Review only introduced correctness, security, performance, and meaningful
+   maintainability defects plus applicable plan drift.
+3. Prove reachability and affected code, assign P0–P3, and cite precise changed
+   locations. Avoid speculation and nits.
+4. Return a patch verdict and repair targets. Never implement a finding.
 
-Find introduced defects with reproduction and reachability. Rank findings. Never implement fixes or repeat an unchanged review as new proof.
+## Return and recovery
 
-## Return
+Return `BLOCK`, `REPAIR_REQUIRED`, `ACCEPT_WITH_FINDINGS`, or `ACCEPT` with
+candidate ref, findings, evidence, verdict, blocker, next action, and receiver.
+Review again only after candidate-changing repairs, at most three rounds.
 
-Handoff fields: `plan_ref`, `role`, `subject`, `depends_on`, `scope`, `authority`, `outcome`, `evidence`, `blocker`, `next_action`, `receiving_role`.
-
-Outcomes: `BLOCK`, `REPAIR_REQUIRED`, `ACCEPT_WITH_FINDINGS`, `ACCEPT`. Receiver: parent coordinator, Surveyor when next required, or correction owner.
-
-## Independence
-
-Fresh non-author only. Never automatic per slice. Candidate authors never supply this verdict.
-
-## Correction
-
-Returns repair targets only. Does not edit the candidate or re-review without new evidence.
+Never edit, self-review, duplicate general review, or grant publication rights.

package/skills/repository-fit/SKILL.md

@@ -1,33 +1,39 @@
 ---
 name: repository-fit
 description: >
-  Propose one target repository and plan-directory assumption when the target
-  is absent or ambiguous. Use for repository fit, choose repo, or confirm
-  workspace before Set Bearings. Do not use after fit is confirmed, for product
-  edits, routing roles, or owner-only publication decisions.
+  Verify and propose one target repository and plan directory when repository
+  identity is absent or ambiguous. Use for Repository Fit before Set Bearings.
+  Do not use after owner-confirmed fit, for parent-wide exploration, planning
+  decisions, product edits, role routing, or publication.
 ---
 
 # Repository Fit
 
-Procedural stage. Not a persona.
+Fresh planning node. It returns evidence; the Router records the decision.
 
-## Match / non-match
+## Inputs and match
 
-- **Match:** selected repository is missing, multi-root, or identity is ambiguous.
-- **Non-match:** owner already confirmed fit; Set Bearings or later stages are ready.
+- **Inputs:** Journey goal, authorized candidate roots, visible prior decisions,
+  repository rules, and discovery limits.
+- **Match:** target root, repository identity, or plan location is unresolved.
+- **Non-match:** the owner already named and confirmed the repository and plan.
 
-## Inputs
+## Algorithm
 
-Work goal, candidate roots, and any prior fit evidence.
+1. Inspect the selected root only. Open one additional root only when authorized.
+2. Cap discovery at depth 4 and 200 paths; prefer Git identity, manifests,
+   top-level instructions, and current plan conventions.
+3. Reject nested-repository confusion, missing Git identity, or conflicting
+   owner evidence with a typed finding.
+4. Propose exactly one repository and one plan-directory assumption with path
+   evidence and a short reason.
+5. Ask one owner confirmation question. A recommendation is not confirmation.
 
-## Procedure
+## Return and recovery
 
-1. Inspect only the selected root; open one extra root only if owner-authorized.
-2. Cap discovery (depth 4, 200 paths). Prefer manifests, top-level docs, plan conventions.
-3. Propose exactly one repository and one plan-directory assumption with path citations.
-4. Ask one concise owner confirmation question. Do not treat a recommendation as approval.
-5. Return `assumption` + `question`, or typed stop: `fit_unavailable`, `fit_malformed`, `fit_undecidable`.
+Return `FIT_PROPOSED`, `FIT_CONFIRMED`, `FIT_UNAVAILABLE`, `FIT_MALFORMED`, or
+`FIT_UNDECIDABLE` with root, plan directory, evidence, blocker, and next action.
+Retry only from new evidence, at most three attempts.
 
-## Never
-
-Write files, walk parents without authority, rank alternatives after undecidable, or advance into Set Bearings.
+Never write files, walk unapproved parents, rank endless alternatives, or
+advance the Journey itself.

package/skills/set-bearings/SKILL.md

@@ -1,33 +1,39 @@
 ---
 name: set-bearings
 description: >
-  Create or resume the bounded plan workspace and repository map when project
-  context or current-state map is missing after fit. Use for set bearings,
-  plan workspace, or repository map. Do not use before fit is confirmed, for
-  owner decisions, design drafts, product edits, or risk-profile authoring.
+  Create or resume the visible Journey workspace and bounded repository map
+  after Repository Fit is confirmed. Use for Set Bearings or stale project
+  context. Do not use before fit, for owner decisions, design, implementation,
+  hidden state, or deleting historical plans.
 ---
 
 # Set Bearings
 
-Procedural stage. Not a persona.
+Fresh planning node. The Router announces `Setting Our Bearings in <repo>.`
 
-## Match / non-match
+## Inputs and match
 
-- **Match:** fit is confirmed and plan workspace or repository map is missing or stale.
-- **Non-match:** workspace already ready; Gather Supplies or later stages apply.
+- **Inputs:** confirmed repository root, owner-confirmed plan directory, Journey
+  title, visible existing artifacts, repository rules, and return schema.
+- **Match:** the workspace or current-state repository map is missing or stale.
+- **Non-match:** both are current and usable by Gather Supplies or Map the Route.
 
-## Inputs
+## Algorithm
 
-Repository root, work goal, and owner-confirmed plan directory (verbatim).
+1. Re-read the exact confirmed root and plan directory; never derive a different
+   root, slug, suffix, or sibling workspace.
+2. Preserve every existing artifact and unrelated edit. Resume rather than
+   replace an existing Journey.
+3. Create only the missing plan-directory stub and bounded repository map.
+4. Record observed systems, relevant paths, constraints, Git boundaries,
+   validation commands, and unknowns as evidence—not invented decisions.
+5. Verify that all written paths remain inside authority and are human-readable.
 
-## Procedure
+## Return and recovery
 
-1. Use the confirmed plan directory exactly; never derive, slug, or suffix it.
-2. Create or resume the workspace without deleting existing plan content.
-3. Write only the plan-directory stub and sibling repository map.
-4. Leave risk-profile content to Gather Supplies / owner; do not invent it.
-5. Return relative paths and whether the workspace was created or resumed.
+Return `WORKSPACE_READY`, `WORKSPACE_RESUMED`, `NEEDS_EVIDENCE`, or `BLOCKED`
+with paths, map freshness, evidence, blocker, and next planning stage. Retry only
+from corrected evidence, at most three attempts.
 
-## Never
-
-Ask planning questions, design, implement, edit product code, or delete plans or hidden runtime state.
+Never ask planning questions, write risk choices, design, implement, create
+hidden runtime state, or delete plan content.

package/skills/sub-explorer/SKILL.md

@@ -1,37 +1,39 @@
 ---
 name: sub-explorer
 description: >
-  Orchestrate bounded packets inside one Explorer-owned wave when at least two
-  proven-independent lanes need one extra coordination level. Use for
-  sub-explorer or in-wave lane coordination. Do not use as a second nesting
-  level, for product implementation, multi-wave control, or assurance.
+  Coordinate at least two proven-independent lanes inside one Explorer wave in
+  a fresh session. Use for Sub-Explorer or necessary in-wave lane coordination.
+  Do not use for one lane, nested Sub-Explorers, implementation, cross-wave
+  control, planning, or assurance.
 ---
 
-# Sub-explorer
+# Sub-Explorer
 
-## Trigger
+Second mutation-authority level; coordinates more and implements less than a
+Crewmate. Maximum one Sub-Explorer level.
 
-- **Match:** one Explorer-owned wave has ≥2 proven-independent lanes and one extra coordination level materially reduces context or conflict.
-- **Non-match:** single lane; nested Sub-explorer requests; Trail Boss multi-wave work; direct Crewmate packets.
+## Inputs and match
 
-## Inputs
+- **Inputs:** parent wave, lane boundaries, packet specs, dependencies,
+  authority, lineup identities, cadence, and return schema.
+- **Match:** two or more disjoint lanes materially reduce context or conflict.
+- **Non-match:** direct packets suffice, independence is unproven, or another
+  Sub-Explorer level is requested.
 
-Parent wave plan path, lane boundaries, packet specs, dependencies, authority, `required_assurance`, expected handoff.
+## Algorithm
 
-## Responsibility
+1. Start fresh; verify parent authority, lane disjointness, dependencies, and
+   approved primary/fallback identity for each node.
+2. Dispatch fresh Crewmate sessions with only their bounded packet and evidence.
+3. Track returns by packet ID; preserve foreign changes and never write a packet.
+4. Reconcile lane evidence and return conflicts to the Explorer.
+5. Request assurance only at the recorded cadence boundary and only through the
+   parent coordinator.
 
-Orchestrate in-wave packets only. Maximum one Sub-explorer level. Never perform packets, expand parent wave authority, or assure results.
+## Return and recovery
 
-## Return
+Return `READY`, `REROUTED`, `WAITING_ON`, or `OWNER_DECISION_REQUIRED` with
+plan ref, role, subject, dependencies, scope, authority, evidence, blocker, next
+action, and receiver. Allow three evidence-changing corrections per lane.
 
-Handoff fields: `plan_ref`, `role`, `subject`, `depends_on`, `scope`, `authority`, `outcome`, `evidence`, `blocker`, `next_action`, `receiving_role`.
-
-Outcomes: `READY`, `REROUTED`, `WAITING_ON`, `OWNER_DECISION_REQUIRED`. Receiver: Crewmate or Explorer.
-
-## Independence
-
-Dormant unless independence is proven. Does not substitute for Explorer, Trail Boss, or Navigator.
-
-## Correction
-
-Return lane conflicts to Explorer. Out-of-wave expansion escalates to parent coordinator.
+Never implement, add nesting, expand the wave, or provide assurance.

package/skills/surveyor/SKILL.md

@@ -1,38 +1,40 @@
 ---
 name: surveyor
 description: >
-  Compare user-facing behavior of a stable integrated candidate with the
-  approved outcome at multi-packet phase, Expedition, Final Audit, or
-  owner-selected acceptance boundaries. Use for surveyor, acceptance review, or
-  outcome comparison. Do not use for implementation, per-packet sufficiency,
-  introduced-defect triage as Park Ranger, or owner-only publication approval.
+  Independently compare one stable integrated candidate with the approved
+  user-facing outcome in a fresh session at the selected acceptance boundary.
+  Use for Surveyor or outcome acceptance. Do not use for implementation,
+  evidence scoring, defect review, automatic review, or publication approval.
 ---
 
 # Surveyor
 
-## Trigger
+Independent outcome assurance, outside the mutation-authority ladder.
 
-- **Match:** multi-packet phase, Expedition, Final Audit, or owner-selected acceptance boundary has a stable integrated candidate.
-- **Non-match:** single unfinished packet; Validator-only sufficiency; Park Ranger defect pass; owner publication decisions reserved to Owner Authority.
+## Inputs and match
 
-## Inputs
+- **Inputs:** approved specification/review baseline, exact integrated candidate,
+  author identities, user-facing evidence, cadence boundary, prior required
+  assurance, and return schema.
+- **Match:** Surveyor is declared and a selected review boundary has an integrated
+  candidate suitable for user-facing comparison.
+- **Non-match:** unfinished packet, unstable candidate, Validator sufficiency,
+  Park Ranger defect review, or owner-only release decision.
 
-Plan path, exact integrated `candidate_ref`, approved outcome, user-facing evidence, prior required assurance results, expected handoff.
+## Algorithm
 
-## Responsibility
+1. Start fresh; verify candidate continuity, independence, completed prerequisite
+   assurance, and the cadence boundary.
+2. Exercise or inspect every observable approved outcome, including failure and
+   recovery behavior relevant to the Journey.
+3. Map each gap to an exact requirement and evidence location. Separate observed
+   behavior from inference.
+4. Return acceptance or gaps without repairing the candidate.
 
-Read-only acceptance comparison against the approved outcome. Report gaps with precise locations. Never repair implementation or approve owner-only actions.
+## Return and recovery
 
-## Return
+Return `ACCEPT`, `GAPS`, or `OWNER_DECISION_REQUIRED` with candidate ref,
+requirement coverage, evidence, gaps, blocker, next action, and receiver.
+Reassess only a new integrated candidate, at most three rounds.
 
-Handoff fields: `plan_ref`, `role`, `subject`, `depends_on`, `scope`, `authority`, `outcome`, `evidence`, `blocker`, `next_action`, `receiving_role`.
-
-Outcomes: `ACCEPT`, `GAPS`, `OWNER_DECISION_REQUIRED`. Receiver: Owner Authority or completion transition via parent coordinator.
-
-## Independence
-
-Fresh non-author only. Distinct from Validator sufficiency and Park Ranger defect review. Authors never self-accept.
-
-## Correction
-
-Returns gaps only. Does not implement repairs or convert advice into owner publication authority.
+Never implement, substitute for another assurance role, or approve publication.

package/skills/trail-boss/SKILL.md

@@ -1,38 +1,40 @@
 ---
 name: trail-boss
 description: >
-  Trail Boss cross-wave coordination for concurrent Explorer-owned waves with
-  shared-interface or dependency conflicts. Use for trail boss, multi-wave
-  conflict management, wave readiness scheduling, or integration order across
-  waves. Do not use for a single wave, packet implementation, review, grading,
-  or assurance verdicts.
+  Coordinate concurrent Explorer waves that share interfaces, dependencies, or
+  integration conflicts in a fresh session. Use for Trail Boss, cross-wave
+  readiness, or integration ordering. Do not use for one wave, packet work,
+  planning, model selection, review, grading, or assurance.
 ---
 
 # Trail Boss
 
-## Trigger
+Cross-wave authority. Coordinates more and implements less than Explorer.
 
-- **Match:** ≥2 Explorer-owned waves are active or need cross-wave dependency, shared-interface, or conflict management.
-- **Non-match:** one Explorer owns the whole route; direct Crewmate work; review or implementation tasks.
+## Inputs and match
 
-## Inputs
+- **Inputs:** approved wave graph, lane status, shared interfaces, dependencies,
+  authority, lineup, cadence, budgets, and return schema.
+- **Match:** at least two active Explorer waves require conflict or integration
+  control.
+- **Non-match:** one Explorer can sequence all work or the request is execution
+  or assurance.
 
-Plan path, wave graph, lane status, budgets, shared interfaces, `required_assurance`, expected handoff.
+## Algorithm
 
-## Responsibility
+1. Start fresh; verify every wave owner, dependency edge, shared interface, and
+   candidate boundary.
+2. Schedule readiness and integration order. Serialize conflicting writes;
+   allow parallel waves only when independence is proven.
+3. Dispatch fresh Explorer sessions with bounded wave packets and named lineup.
+4. Reconcile typed returns and surface interface drift without repairing it.
+5. Tell the Navigator when the cadence boundary is ready for assurance; never
+   create an extra review round.
 
-Schedule wave readiness and integration order only. Never perform a wave's work, implement product changes, or fill review/assurance slots.
+## Return and recovery
 
-## Return
+Return `READY`, `REROUTED`, `WAITING_ON`, or `OWNER_DECISION_REQUIRED` with
+plan ref, role, subject, dependencies, scope, authority, evidence, blocker, next
+action, and receiver. Try at most three evidence-changing schedules.
 
-Handoff fields: `plan_ref`, `role`, `subject`, `depends_on`, `scope`, `authority`, `outcome`, `evidence`, `blocker`, `next_action`, `receiving_role`.
-
-Outcomes: `READY`, `REROUTED`, `WAITING_ON`, `OWNER_DECISION_REQUIRED`. Receiver: Explorer, Navigator, or assurance role when required.
-
-## Independence
-
-Orchestration-only. Omitted when one Explorer suffices. Never self-certifies integrated candidates.
-
-## Correction
-
-Escalate unresolvable cross-wave conflicts to Navigator or Owner Authority; do not absorb Explorer packet work.
+Never implement, absorb Explorer work, self-assure, or expand authority.