@voybio/ace-swarm 0.1.0

package/assets/.agents/skills/schema-forge/SKILL.md

@@ -0,0 +1,174 @@
+---
+name: schema-forge
+description:
+  Evolve ACE interface schemas safely with versioning, compatibility analysis, and migration guidance. Use when handoff, status, or artifact contracts change or a schema violation must be repaired.
+---
+
+# Schema Forge
+
+## Purpose
+
+Guarantee interoperability by evolving schemas with strict validation, versioning, and consumer-safe migration rules.
+Schema without versioning is a landmine. Schema with versioning is a contract.
+
+## Canonical Use Cases
+
+1. A handoff, status event, or artifact schema needs a new field or compatibility-safe evolution.
+2. A breaking contract change is proposed and the team needs versioning, migration guidance, and producer/consumer mapping.
+3. A schema violation exposed drift between contract and consumers, and the interface must be repaired safely.
+
+---
+
+## MICE Boundaries
+
+| Rule | Enforcement |
+|---|---|
+| **Modular** | Manages schema contracts; does NOT implement business logic or change routing. |
+| **Interoperable** | Every schema follows JSON Schema draft-07+ with `$id` and `$schema` fields. |
+| **Customizable** | Schema evolution respects `TEAL_CONFIG.md` consumer topology. |
+| **Extensible** | New schemas are added with producer/consumer mapping, not ad-hoc. |
+
+---
+
+## MICE/TEAL Alignment
+
+- Interoperable-first skill: schema contracts are the module boundary glue.
+- TEAL-aware: schema evolution must preserve producer/consumer compatibility across configured routes.
+
+---
+
+## Trigger Conditions
+
+Use when:
+
+- adding/changing handoff fields
+- adding/changing status event payloads
+- introducing new artifact manifests
+- upgrading module interface versions
+- post-incident contract repair
+- any agent reports `schema_violation` via Wrong-Stuff Protocol
+
+---
+
+## Inputs
+
+- `agent-state/MODULES/schemas/*.json`
+- `agent-state/INTERFACE_REGISTRY.md`
+- `agent-state/TEAL_CONFIG.md`
+- producer/consumer role manifests
+- the proposed change (what fields are being added/changed/removed)
+
+---
+
+## Socratic Questions (Pre-Evolution)
+
+- Which invariant is this schema enforcing?
+- Who produces it and who consumes it?
+- Is this additive or breaking for current consumers?
+- What migration path is required?
+- Can consumers tolerate this change without code modification?
+
+---
+
+## Version Bump Decision Matrix
+
+| Change Type | Examples | Version Bump | Consumer Impact |
+|---|---|---|---|
+| New optional field | Adding `metadata.tags` | `patch` | None; consumers ignore unknown fields |
+| New required field | Adding `context.priority` | `minor` | Consumers must handle; provide default |
+| Field type change | String → Object | `major` | Breaking; migration required |
+| Field removal | Removing `legacy_id` | `major` | Breaking; migration required |
+| Enum value addition | Adding `ROLLBACK` to decision types | `minor` | Consumers must handle unknown enum gracefully |
+| Enum value removal | Removing deprecated value | `major` | Breaking; migration required |
+| Constraint tightening | maxLength 100 → 50 | `major` | Breaking for producers |
+| Constraint relaxation | maxLength 50 → 100 | `patch` | None |
+
+---
+
+## Compatibility Rules
+
+| Rule | Enforcement |
+|---|---|
+| **Backward Compatible** | Consumers on version N must still parse payloads from version N+1 (additive only). |
+| **Forward Compatible** | Producers on version N+1 must still produce valid payloads for version N consumers (no removal). |
+| **Breaking Change** | Requires: migration guide, consumer notification, deprecation period (>= 1 sprint), evidence of consumer readiness. |
+| **Schema Coexistence** | During migration, both old and new schema versions must be loadable simultaneously. |
+
+---
+
+## Workflow — Mapped to Clarity Protocol
+
+### `[STATE_ANALYSIS]`
+1. Identify the schema being evolved and its current version.
+2. Map all producers and consumers from `INTERFACE_REGISTRY.md`.
+3. Assess the proposed change against compatibility rules.
+
+### `[STRATEGY_SELECTOR]`
+4. Determine version bump using decision matrix.
+5. If `major`: plan migration guide, deprecation period, consumer notification.
+6. If `minor/patch`: plan additive change with backward compatibility preserved.
+
+### `[EXECUTION_LOG]`
+7. Update schema file with version bump and change.
+8. Add `$id` with new version number.
+9. For breaking changes: write migration guide with before/after examples.
+10. Validate updated schema against sample payloads (both old and new format).
+11. Update `INTERFACE_REGISTRY.md` with new version and consumer mapping.
+
+### `[ARTIFACT_UPDATE]`
+12. Write updated schema to `agent-state/MODULES/schemas/<name>.json`.
+13. Update `agent-state/INTERFACE_REGISTRY.md` entry.
+14. Write `agent-state/SCHEMA_CHANGELOG.md` entry:
+
+```markdown
+### <version> — <ISO8601>
+- **Schema:** <name>
+- **Change:** <description>
+- **Bump:** <patch | minor | major>
+- **Compatibility:** <backward | forward | breaking>
+- **Migration:** <N/A | see migration guide>
+- **Producers:** <list>
+- **Consumers:** <list>
+- **Validated:** <sample payloads tested>
+```
+
+15. Append evidence to `EVIDENCE_LOG.md`.
+
+### `[VERIFICATION]`
+16. Schema validates against JSON Schema meta-schema.
+17. Sample payloads (old + new format) validate against updated schema.
+18. Version bump matches change type per decision matrix.
+19. All producers and consumers are mapped in `INTERFACE_REGISTRY.md`.
+20. Breaking changes have migration guide and deprecation plan.
+
+---
+
+## Output Contract
+
+- Updated schema file(s) under `agent-state/MODULES/schemas/`
+- Updated `agent-state/INTERFACE_REGISTRY.md`
+- `agent-state/SCHEMA_CHANGELOG.md` entry
+
+---
+
+## Anti-Patterns
+
+| Anti-Pattern | Correct Behavior |
+|---|---|
+| Changing schema without version bump | Every change gets a version bump per decision matrix |
+| Breaking change without migration guide | Major bumps require migration guide with examples |
+| Schema without producer/consumer mapping | Every schema has explicit ownership in `INTERFACE_REGISTRY.md` |
+| "Just add the field" | Assess compatibility impact before adding |
+| Schema without sample payloads | Every schema version has validated sample payloads |
+
+---
+
+## Failure Policy
+
+Reject schema updates that:
+
+- lack version updates
+- omit compatibility notes
+- have no producer/consumer mapping
+- fail sample payload validation
+- introduce breaking changes without migration guide

package/assets/.agents/skills/skill-auditor/SKILL.md

@@ -0,0 +1,52 @@
+---
+name: skill-auditor
+description: 
+  Audit discoverable ACE skills against the portable skill contract, validation expectations, and progressive-disclosure rules. Use when the user asks which skills are production-ready, what is missing from the skill catalog, or how to verify portability across Claude, Cursor, Codex, Antigravity, and similar clients.
+---
+
+# Skill Auditor
+
+Use this skill to inspect the current ACE skill catalog before a remediation pass, release gate, or new skill rollout. Its job is to turn vague quality concerns into a concrete audit report and prioritized backlog.
+
+## Canonical Use Cases
+
+1. A user asks which existing skills are not yet production-ready.
+2. The team needs a readiness rubric before hardening trigger metadata or validation canaries.
+3. A portability change lands and someone needs to verify that `SKILL.md` remains canonical while any client overlays stay optional.
+
+## Inputs
+
+- The discoverable skill catalog from workspace and package-default skills.
+- Each skill's `SKILL.md`, optional `agents/` overlays, and supporting `references/` directory if present.
+- The current readiness criteria for trigger descriptions, core sections, validation, and portability.
+
+## Workflow
+
+1. Enumerate the discoverable skills and inspect each portable skill contract.
+2. Check frontmatter, trigger-aware description quality, canonical use cases, inputs/workflow/outputs/failure sections, validation surface, portability notes, and progressive-disclosure discipline.
+3. Record per-skill findings as pass, warn, or fail, keeping optional adapter overlays as neutral inventory rather than required readiness.
+4. Write `agent-state/SKILL_AUDIT_REPORT.md` with summary counts, per-skill findings, and rollout backlog items.
+
+## Outputs
+
+- `agent-state/SKILL_AUDIT_REPORT.md`
+- Per-skill pass/warn/fail findings with concrete remediation notes
+- A rollout backlog for validation, portability, and progressive-disclosure gaps
+
+## Validation
+
+- Run the `skill_audit` MCP tool when it is available, or perform a direct catalog inspection of discoverable `SKILL.md` files, and verify that `agent-state/SKILL_AUDIT_REPORT.md` is written.
+- Confirm that a skill missing required sections is flagged as `fail`.
+- Confirm that a skill with optional client overlays but no portability note is flagged as `warn`, not as a hard failure for missing `openai.yaml`.
+
+## Compatibility
+
+- `SKILL.md` is the canonical source of truth for the audit.
+- Files under `agents/` are optional overlays only and should be audited for parity, not treated as mandatory.
+- The audit must stay meaningful even if no client-specific adapters exist at all.
+
+## Failure Policy
+
+- If the catalog cannot be read, stop and surface the specific filesystem problem.
+- If a skill is ambiguous, lower confidence through `warn` findings instead of silently passing it.
+- Do not treat the absence of a specific client adapter file as a failure unless the portable contract itself depends on it.

package/assets/.agents/skills/state-auditor/SKILL.md

@@ -0,0 +1,182 @@
+---
+name: state-auditor
+description:
+  Deterministic integrity scanner for ACE state artifacts. Use before/after role transitions, release gates, or when status truth may have drifted.
+---
+
+# State Auditor
+
+## Purpose
+
+Detect and repair state drift across `agent-state/` so autonomous execution remains coherent and reproducible.
+State without audit is faith. State with audit is engineering.
+
+## Canonical Use Cases
+
+1. A role transition or handoff is about to occur and the team needs a fast check that core state files agree.
+2. A blocker persists or contradictory state is suspected, and someone needs a deterministic drift report before acting.
+3. A release or post-incident checkpoint needs proof that status, evidence, risks, and routing state are coherent.
+
+---
+
+## MICE Boundaries
+
+| Rule | Enforcement |
+|---|---|
+| **Modular** | Audits state consistency; does NOT implement features or change scope. |
+| **Interoperable** | Audit reports follow canonical schema below. |
+| **Customizable** | Recency thresholds adapt to sprint cadence in `TEAL_CONFIG.md`. |
+| **Extensible** | New audit checks added as checklist rows, not prose. |
+
+---
+
+## MICE/TEAL Alignment
+
+- Modular: audits state only, does not implement feature logic.
+- TEAL-aware: validates status/handoff consistency against active pipeline topology.
+
+---
+
+## Trigger Conditions
+
+Use when:
+
+- a role transition is about to occur
+- handoff payload changed
+- blockers persist > 1 cycle
+- there is contradiction between `STATUS.md` and `EVIDENCE_LOG.md`
+- pre-release gate check
+- post-incident state verification
+
+---
+
+## Inputs
+
+- `agent-state/TASK.md`
+- `agent-state/STATUS.md`
+- `agent-state/SCOPE.md`
+- `agent-state/DECISIONS.md`
+- `agent-state/RISKS.md`
+- `agent-state/EVIDENCE_LOG.md`
+- `agent-state/HANDOFF.json`
+- `agent-state/TEAL_CONFIG.md`
+
+---
+
+## Audit Checklist
+
+| # | Check | Pass Criteria | On Fail |
+|---|---|---|---|
+| 1 | **Artifact Presence** | All 8 required inputs exist on disk | CRITICAL: list missing files |
+| 2 | **Artifact Recency** | All files modified within current sprint window | WARN if > 1 sprint; CRITICAL if > 2 |
+| 3 | **Owner Consistency** | Current owner in `STATUS.md` matches `HANDOFF.json` `to` field | CRITICAL: contradictory ownership |
+| 4 | **Phase Consistency** | Phase in `STATUS.md` matches `HANDOFF.json` context | WARN: possible stale handoff |
+| 5 | **Evidence Anchors** | Every claimed transition in `STATUS.md` has `EVIDENCE_LOG.md` entry | CRITICAL: unanchored transitions |
+| 6 | **Blocker Resolution** | Active blockers in `STATUS.md` have owner and due condition | WARN: orphaned blockers |
+| 7 | **Decision Currency** | `DECISIONS.md` entries reference current (not superseded) evidence | WARN: stale decisions |
+| 8 | **Risk Currency** | `RISKS.md` entries scored within current sprint | WARN: unscored risks |
+| 9 | **Scope Alignment** | `SCOPE.md` boundaries not contradicted by `TASK.md` objectives | CRITICAL: scope/task misalignment |
+| 10 | **TEAL Route Validity** | Configured routes in `TEAL_CONFIG.md` match actual agent topology | CRITICAL: invalid routes |
+
+---
+
+## Contradiction Detection Protocol
+
+When contradictions are found between state files:
+
+1. **Identify**: Which two files contain conflicting claims.
+2. **Timestamp**: Which claim is more recent (evidence-linked).
+3. **Classify**: Is this a stale-state issue or a genuine disagreement.
+4. **Route**: Stale → update the older file. Genuine → emit `FIX_REQUIRED` to owner.
+5. **Record**: Log contradiction and resolution in audit report.
+
+---
+
+## Workflow — Mapped to Clarity Protocol
+
+### `[STATE_ANALYSIS]`
+1. Check all 8 required artifacts for presence and recency.
+2. Read `STATUS.md` current owner and phase.
+3. Read `HANDOFF.json` for claimed transition state.
+
+### `[STRATEGY_SELECTOR]`
+4. Run full audit checklist (10 checks) in order.
+5. Classify each result as `PASS`, `WARN`, or `CRITICAL`.
+6. If any CRITICAL found, plan immediate routing.
+
+### `[EXECUTION_LOG]`
+7. Execute each audit check deterministically.
+8. Cross-check owner/phase consistency between `STATUS.md` and handoff.
+9. Verify evidence anchors exist for each claimed transition.
+10. Detect contradictory statements across core state files using contradiction protocol.
+11. Validate TEAL route configuration against actual topology.
+
+### `[ARTIFACT_UPDATE]`
+12. Write `agent-state/STATE_AUDIT_REPORT.md`:
+
+```markdown
+# State Audit Report
+Generated: <ISO8601>
+Overall: <PASS | WARN | CRITICAL>
+
+## Checklist Results
+| # | Check | Status | Detail |
+|---|---|---|---|
+| 1 | Artifact Presence | PASS/WARN/CRITICAL | <detail> |
+| ... | ... | ... | ... |
+
+## Critical Findings
+- <finding with evidence ref and required action>
+
+## Non-Blocking Findings
+- <finding with evidence ref>
+
+## Contradictions Detected
+- <file1> vs <file2>: <description> → <resolution>
+
+## Required Actions
+| Action | Owner | Due Condition |
+|---|---|---|
+| <action> | <agent-role> | <what must happen> |
+```
+
+13. Append audit evidence to `EVIDENCE_LOG.md`.
+
+### `[VERIFICATION]`
+14. Every checklist row has a deterministic result.
+15. All CRITICAL findings have owner + required action.
+16. Contradictions are logged with both references.
+17. Report is internally consistent (no contradictions within the report itself).
+
+---
+
+## Output Contract
+
+- `agent-state/STATE_AUDIT_REPORT.md`
+  - `summary` (overall PASS/WARN/CRITICAL)
+  - `critical_findings` (with evidence refs and required actions)
+  - `non_blocking_findings` (with evidence refs)
+  - `contradictions` (with dual refs and resolution)
+  - `required_actions` (with owner and due condition)
+
+---
+
+## Anti-Patterns
+
+| Anti-Pattern | Correct Behavior |
+|---|---|
+| "State looks fine" | Every check produces PASS/WARN/CRITICAL with evidence |
+| Skipping contradiction detection | Always scan for cross-file contradictions |
+| Audit without artifact presence check | Presence check is always step 1 |
+| Reporting WARN but not tracking | All WARNs escalate to CRITICAL after 2 consecutive audits |
+
+---
+
+## Failure Policy
+
+If critical contradiction is found:
+
+- emit `GATE_FAILED` with `failure_type: state_drift`
+- block downstream transitions
+- assign owner and due condition in `STATUS.md`
+- if > 3 CRITICAL findings in one audit, escalate to incident-commander

package/assets/.github/hooks/ace-copilot.json

@@ -0,0 +1,68 @@
+{
+  "hooks": {
+    "SessionStart": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "UserPromptSubmit": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "PreToolUse": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "PostToolUse": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "PreCompact": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "SubagentStart": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "SubagentStop": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ],
+    "Stop": [
+      {
+        "type": "command",
+        "command": "node ./.agents/ACE/scripts/ace/ace-hook-dispatch.mjs",
+        "windows": "node .\\\\.agents\\\\ACE\\\\scripts\\\\ace\\\\ace-hook-dispatch.mjs",
+        "timeout": 10
+      }
+    ]
+  }
+}

package/assets/agent-state/ACE_WORKFLOW.md

@@ -0,0 +1,131 @@
+---
+ace_runtime_version: "1.0.0"
+runtime:
+  mode: "interactive"
+  profile_name: "default"
+workspace:
+  root: ".agents/ACE/.ace/workspaces"
+  retention: "delete"
+  hooks:
+    after_create: null
+    before_run: null
+    after_run: null
+    before_remove: null
+executor:
+  command: null
+  approval_policy: "on-request"
+  thread_sandbox: "workspace-write"
+  turn_sandbox_policy: null
+  max_turns: 6
+  turn_timeout_ms: 300000
+tools:
+  registry_path: "agent-state/runtime-tool-specs.json"
+autonomy:
+  orchestrator_preflight: true
+  recall_context: true
+  state_sources: ["agent-state/TASK.md", "agent-state/SCOPE.md", "agent-state/QUALITY_GATES.md", "agent-state/STATUS.md", "agent-state/HANDOFF.json", "agent-state/EVIDENCE_LOG.md"]
+  before_run_checks: []
+  stop_checks: []
+  review_mode: null
+continuity:
+  enabled: true
+  source_paths: ["agent-state/TASK.md", "agent-state/SCOPE.md", "agent-state/QUALITY_GATES.md", "agent-state/HANDOFF.json", "agent-state/STATUS.md", "agent-state/DECISIONS.md", "agent-state/RISKS.md", "agent-state/EVIDENCE_LOG.md"]
+  max_sources: 6
+  max_excerpt_chars: 450
+  max_total_chars: 2600
+  max_status_events: 6
+  max_run_ledger_entries: 4
+  include_snapshot: true
+tracker:
+  kind: "none"
+  config: {}
+observability:
+  http_enabled: false
+  http_port: null
+  vericify_bridge:
+    enabled: false
+    bridge_path: "agent-state/vericify/ace-bridge.json"
+    process_post_path: "agent-state/vericify/process-posts.json"
+---
+# ACE Runtime Prompt Template
+
+Operate under ACE runtime profile `{{runtime.profile_name}}` in `{{runtime.mode}}` mode.
+
+## Runtime Intent
+- Use this profile for ACE execution that must stay aligned to file-backed state and managed workspace policy.
+- Treat front matter as execution policy. If a field is `null`, `false`, or empty, do not invent extra runtime machinery.
+- Preserve objective, boundary, evidence, and transition truth inside ACE artifacts rather than chat-local memory.
+
+## Task
+{{task}}
+
+## Workspace
+- Managed root: `{{workspace.root}}`
+- Retention: `{{workspace.retention}}`
+
+## Executor
+- Command: `{{executor.command}}`
+- Approval policy: `{{executor.approval_policy}}`
+- Thread sandbox: `{{executor.thread_sandbox}}`
+- Max turns: `{{executor.max_turns}}`
+- Turn timeout ms: `{{executor.turn_timeout_ms}}`
+
+## Runtime Tools
+- Registry: `{{tools.registry_path}}`
+
+## Autonomy
+- Orchestrator preflight: `{{autonomy.orchestrator_preflight}}`
+- Recall context: `{{autonomy.recall_context}}`
+- State sources: `{{autonomy.state_sources}}`
+- Before-run checks: `{{autonomy.before_run_checks}}`
+- Stop checks: `{{autonomy.stop_checks}}`
+- Review mode: `{{autonomy.review_mode}}`
+
+## Autonomy Guidance
+- `orchestrator_preflight` means validate ACE truth before substantial work or delegation.
+- `recall_context` means prefer current task, status, handoff, and evidence state over stale thread assumptions.
+- `before_run_checks` are blocking checks. If they fail, stop and surface the failure through ACE evidence and status channels.
+- `stop_checks` are terminal or handoff checks. If they fail, do not continue optimistically.
+- `review_mode` must stay inside existing ACE tools and artifacts; it is an overlay, not a second subsystem.
+
+## Continuity
+- Enabled: `{{continuity.enabled}}`
+- Source paths: `{{continuity.source_paths}}`
+- Max sources: `{{continuity.max_sources}}`
+- Max excerpt chars: `{{continuity.max_excerpt_chars}}`
+- Max total chars: `{{continuity.max_total_chars}}`
+- Max status events: `{{continuity.max_status_events}}`
+- Max run-ledger entries: `{{continuity.max_run_ledger_entries}}`
+- Include snapshot: `{{continuity.include_snapshot}}`
+
+## Continuity Guidance
+- Continuity packets are derived from current ACE truth; they are not a replacement source of truth.
+- Treat the configured budgets as source-side compaction rules for unattended execution and delegation, including recent status and run-ledger visibility.
+- If continuity is disabled, do not synthesize a replacement packet in chat or hidden state.
+
+## ACE Recall
+{{ace_state_recall_md}}
+
+## Context Snapshot
+{{ace_context_snapshot_md}}
+
+## Continuity Packet
+{{ace_continuity_packet_md}}
+
+## Tracker
+- Kind: `{{tracker.kind}}`
+
+## Sidecars
+- Vericify bridge enabled: `{{observability.vericify_bridge.enabled}}`
+- Vericify bridge path: `{{observability.vericify_bridge.bridge_path}}`
+- Vericify process-post path: `{{observability.vericify_bridge.process_post_path}}`
+
+## Notes
+- Preserve ACE state artifacts as the durable source of truth.
+- Treat Vericify as an optional sidecar, not a required ACE runtime dependency.
+- Treat this template as additive execution guidance for future unattended flows.
+
+## Failure Handling
+- When the task contract, evidence, or hooks are contradictory, prefer hold/escalation over optimistic continuation.
+- Record blockers in ACE artifacts with owners and evidence refs before handing off or stopping.
+- If the profile is missing capabilities, keep the gap explicit rather than silently broadening the runtime contract.

package/assets/agent-state/ARTIFACT_MANIFEST.json

@@ -0,0 +1,5 @@
+{
+  "schema_version": "1.1.0",
+  "updated_at": "1970-01-01T00:00:00.000Z",
+  "artifacts": []
+}