kiro-spec-engine 1.47.4 → 1.47.6

package/CHANGELOG.md

@@ -8,6 +8,24 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 
 ### Added
+- **Scene ontology impact/path analysis commands**: Added `kse scene ontology impact` (reverse dependency blast-radius analysis with `--relation` and `--max-depth`) and `kse scene ontology path` (shortest relation path between refs with optional `--undirected`) to improve ontology-driven change planning and explainability.
+- **Close-loop quantitative DoD gates**: Added `--dod-max-risk-level`, `--dod-kpi-min-completion-rate`, `--dod-max-success-rate-drop`, and `--dod-baseline-window` so autonomous closure can enforce explicit risk/KPI/baseline thresholds beyond binary checks.
+- **Close-loop conflict/ontology execution planning**: Added lease-conflict scheduling governance and scene ontology scheduling guidance with opt-out controls (`--no-conflict-governance`, `--no-ontology-guidance`), and surfaced planning telemetry in `portfolio.execution_plan` plus agent sync plan output.
+- **Close-loop strategy memory feedback loop**: Added persisted strategy memory (`.kiro/auto/close-loop-strategy-memory.json`) to reuse prior goal-level policy hints and track-level feedback bias during decomposition, with run telemetry under `strategy_memory`.
+- **Unified autonomous observability snapshot**: Added `kse auto observability snapshot` to aggregate close-loop, batch, controller, governance, and KPI-trend telemetry into one machine-readable payload.
+- **Agent-facing spec JSON interfaces**: Added `kse auto spec status <spec-name> --json` and `kse auto spec instructions <spec-name> --json` for structured status/instruction retrieval in master-sub agent workflows.
+- **Autonomous archive schema compatibility tooling**: Added `kse auto schema check` and `kse auto schema migrate` (dry-run by default, `--apply` to persist) to audit and migrate `schema_version` across autonomous archives.
+- **Governance resumed-session filtering and ratio telemetry**: `kse auto governance session list|stats` now support `--resume-only`, with new resumed/fresh counters and resume lineage composition telemetry (`resumed_rate_percent`, `resumed_from_counts`) for resumed-chain observability.
+- **Governance close-loop post-run retention hook**: Added `--governance-session-keep` and `--governance-session-older-than-days` to `kse auto governance close-loop` for automatic governance session archive pruning after each run, with current session protection and telemetry in `governance_session_prune`.
+- **Governance resume drift guardrails**: `kse auto governance close-loop --governance-resume` now reuses persisted target/advisory policy defaults by default and blocks explicit policy drift unless `--governance-resume-allow-drift` is set, reducing accidental resume misconfiguration.
+- **Governance session maintenance commands**: Added `kse auto governance session list|stats|prune` for persisted governance close-loop session observability and retention management, including status/day filters, convergence/risk telemetry, and archive cleanup controls (`--keep`, `--older-than-days`, `--dry-run`).
+- **Governance close-loop session persistence and resume**: Added default governance loop session archiving (`.kiro/auto/governance-close-loop-sessions`) with controls `--governance-session-id`, `--no-governance-session`, and `--governance-resume <session|latest|file>` so interrupted governance rounds can be resumed without restarting from round 1.
+- **Governance close-loop advisory execution**: Added optional advisory action execution to `kse auto governance close-loop` via `--execute-advisory`, including bounded controls `--advisory-recover-max-rounds` and `--advisory-controller-max-cycles`, with per-round advisory telemetry (`advisory_actions`, advisory action counts) and top-level advisory policy/summary output. Advisory source selection is now autonomous (latest recoverable batch summary / latest controller session with pending goals) and emits `skipped` instead of hard-failing when no actionable advisory source exists.
+- **Governance close-loop rounds command**: Added `kse auto governance close-loop` to run bounded governance-maintenance rounds toward a target risk (`--target-risk`) with convergence telemetry (`rounds`, `initial_assessment`, `final_assessment`, `stop_reason`, `converged`) and non-mutating planning support (`--plan-only`, `--dry-run`).
+- **Governance maintenance close-loop command**: Added `kse auto governance maintain` to combine governance assessment and optional maintenance execution (`--apply`) across session/batch/controller archives and recovery memory, with policy knobs (`--session-keep`, `--batch-session-keep`, `--controller-session-keep`, `--recovery-memory-older-than-days`), dry-run support, and before/after telemetry.
+- **Cross-archive governance stats command**: Added `kse auto governance stats` with optional `--days` / `--status` filters to unify session/batch-session/controller-session telemetry and recovery-memory state into one governance snapshot (`totals`, `throughput`, `health` risk diagnostics, `top_master_specs`, and per-archive detail payloads).
+- **Session and batch session stats commands**: Added `kse auto session stats` and `kse auto batch-session stats` with optional `--days` / `--status` filters and JSON telemetry (completion/failure rates, volume aggregates, composition breakdowns, latest-session snapshots) for autonomous archive observability.
+- **Controller session stats command**: Added `kse auto controller-session stats` with optional `--days` / `--status` filters and JSON telemetry (`status_counts`, `queue_format_counts`, completion/failure rates, goal-volume sums, and latest-session snapshot) for controller archive observability.
 - **Session list status filtering + composition telemetry**: `kse auto session list`, `kse auto batch-session list`, and `kse auto controller-session list` now support `--status <csv>` (case-insensitive) and emit `status_filter` / `status_counts` in JSON output for faster archive triage.
 - **KPI mode composition telemetry**: `kse auto kpi trend` JSON output now includes `mode_breakdown` (batch/program/recover/controller/other) to expose mixed-run trend composition.
 - **Spec protection source expansion**: `spec-session prune` now protects specs referenced by recent/incomplete controller sessions (via nested batch summary references), reducing accidental cleanup during controller-driven autonomous runs.

package/README.md

@@ -537,6 +537,11 @@ kse auto close-loop-controller .kiro/auto/program-queue.lines --wait-on-empty --
 kse auto close-loop-controller --controller-resume latest --json # Resume autonomous controller from latest persisted checkpoint
 kse auto controller-session list --limit 50 --json # Inspect persisted close-loop-controller summary sessions
 kse auto controller-session prune --keep 20 --older-than-days 14 --dry-run --json # Prune old controller summaries by retention policy
+kse auto observability snapshot --days 14 --json # Unified autonomous observability snapshot (sessions + governance + trend)
+kse auto spec status 121-00-master --json # Agent-facing structured spec status interface
+kse auto spec instructions 121-02-sub-track --json # Agent-facing execution instructions for one spec
+kse auto schema check --json # Check autonomous archive schema compatibility
+kse auto schema migrate --apply --json # Backfill/migrate schema_version for autonomous archives
 
 # Spec workflow (recommended)
 kse spec bootstrap --name <spec> --non-interactive         # Generate requirements/design/tasks draft
@@ -609,6 +614,8 @@ kse scene contribute --package <path>          # One-stop validate → lint →
 # Scene ontology (NEW in v1.42.0)
 kse scene ontology show --package <path>       # Show ontology graph
 kse scene ontology deps --ref <ref>            # Query dependency chain
+kse scene ontology impact --ref <ref>          # Analyze reverse impact radius
+kse scene ontology path --from <ref> --to <ref> # Query shortest ontology relation path
 kse scene ontology validate --package <path>   # Validate graph consistency
 kse scene ontology actions --ref <ref>         # Show action abstraction
 kse scene ontology lineage --ref <ref>         # Show data lineage

package/README.zh.md

@@ -454,6 +454,11 @@ kse auto close-loop-controller .kiro/auto/program-queue.lines --wait-on-empty --
 kse auto close-loop-controller --controller-resume latest --json # 从最近一次 controller 检查点恢复自治推进
 kse auto controller-session list --limit 50 --json # 查看持久化 close-loop-controller 摘要会话
 kse auto controller-session prune --keep 20 --older-than-days 14 --dry-run --json # 按保留策略清理旧 controller 摘要
+kse auto observability snapshot --days 14 --json # 统一自治可观测快照（会话 + 治理 + 趋势）
+kse auto spec status 121-00-master --json # 面向 Agent 的结构化 Spec 状态接口
+kse auto spec instructions 121-02-sub-track --json # 面向 Agent 的 Spec 执行指令接口
+kse auto schema check --json # 检查自治归档 schema 兼容性
+kse auto schema migrate --apply --json # 回填/迁移自治归档 schema_version
 
 # Spec 工作流（推荐）
 kse spec bootstrap --name <spec> --non-interactive          # 生成 requirements/design/tasks 初稿
@@ -511,6 +516,8 @@ kse scene contribute --package <path>          # 一站式验证 → lint →
 # 场景 Ontology (v1.42.0 新增)
 kse scene ontology show --package <path>       # 显示 Ontology 图
 kse scene ontology deps --ref <ref>            # 查询依赖链
+kse scene ontology impact --ref <ref>          # 分析反向影响面
+kse scene ontology path --from <ref> --to <ref> # 查询最短关系路径
 kse scene ontology validate --package <path>   # 验证图一致性
 kse scene ontology actions --ref <ref>         # 显示 Action Abstraction
 kse scene ontology lineage --ref <ref>         # 显示数据血缘

package/docs/autonomous-control-guide.md

@@ -187,8 +187,10 @@ Session persistence and resume:
 - `--session-keep <n>`: automatically prune old snapshots and keep newest `n` after each close-loop run
 - `--session-older-than-days <n>`: when pruning, only delete snapshots older than `n` days
 - `kse auto session list [--limit <n>] [--status <csv>]`: inspect persisted sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto session stats [--days <n>] [--status <csv>] [--json]`: aggregate session status/topology telemetry in an optional recent-day window
 - `kse auto session prune --keep <n> [--older-than-days <n>]`: enforce retention policy
-  - JSON output includes `status_filter` and `status_counts` for filtered session distributions.
+  - List JSON output includes `status_filter` and `status_counts` for filtered session distributions.
+  - Stats JSON output includes `criteria`, completion/failure rates, `sub_spec_count_sum`, `master_spec_counts`, and `latest_sessions`.
 
 Spec directory maintenance:
 - `kse auto spec-session list [--limit <n>] [--json]`: inspect spec directory inventory under `.kiro/specs`
@@ -203,8 +205,10 @@ Batch summary session persistence and maintenance:
 - `--batch-session-older-than-days <n>`: when pruning, only delete summaries older than `n` days
 - `--no-batch-session`: disable batch summary archive for the current run
 - `kse auto batch-session list [--limit <n>] [--status <csv>]`: inspect persisted batch summary sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto batch-session stats [--days <n>] [--status <csv>] [--json]`: aggregate batch session status/goal-volume telemetry in an optional recent-day window
 - `kse auto batch-session prune --keep <n> [--older-than-days <n>]`: enforce batch summary retention policy
-  - JSON output includes `status_filter` and `status_counts` for filtered status composition.
+  - List JSON output includes `status_filter` and `status_counts` for filtered status composition.
+  - Stats JSON output includes `criteria`, completion/failure rates, goal-volume sums, processed ratio, and `latest_sessions`.
 
 Controller summary session persistence and maintenance:
 - Default controller summary archive: `.kiro/auto/close-loop-controller-sessions/*.json`
@@ -213,8 +217,28 @@ Controller summary session persistence and maintenance:
 - `--controller-session-older-than-days <n>`: when pruning, only delete summaries older than `n` days
 - `--no-controller-session`: disable controller summary archive for the current run
 - `kse auto controller-session list [--limit <n>] [--status <csv>]`: inspect persisted controller summary sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto controller-session stats [--days <n>] [--status <csv>] [--json]`: aggregate controller session status/throughput telemetry in an optional recent-day window
 - `kse auto controller-session prune --keep <n> [--older-than-days <n>]`: enforce controller summary retention policy
-  - JSON output includes `status_filter` and `status_counts` for filtered status composition.
+  - List JSON output includes `status_filter` and `status_counts` for filtered status composition.
+  - Stats JSON output includes `criteria`, `status_counts`, `queue_format_counts`, completion/failure rates, goal-volume sums, and `latest_sessions`.
+
+Cross-archive governance snapshot:
+- `kse auto governance stats [--days <n>] [--status <csv>] [--json]`: aggregate a unified governance cockpit over session/batch/controller archives plus recovery memory state.
+  - JSON output includes `totals`, `throughput`, `top_master_specs`, per-archive `archives.*` stats payloads, and `health` diagnostics (`risk_level`, `concerns`, `recommendations`) for one-command close-loop governance checks.
+- `kse auto governance maintain [--days <n>] [--status <csv>] [--session-keep <n>] [--batch-session-keep <n>] [--controller-session-keep <n>] [--recovery-memory-older-than-days <n>] [--apply] [--dry-run] [--json]`: close-loop governance maintenance entrypoint.
+  - Default is plan-only (`assessment` + `plan`).
+  - `--apply` executes maintenance actions for archive hygiene (session/batch/controller pruning + recovery-memory stale-entry pruning).
+  - `--dry-run` can be combined with `--apply` to validate maintenance impact before deletion.
+- `kse auto governance close-loop [--max-rounds <n>] [--target-risk <low|medium|high>] [--governance-resume <session|latest|file>] [--governance-resume-allow-drift] [--governance-session-id <id>] [--no-governance-session] [--governance-session-keep <n>] [--governance-session-older-than-days <n>] [--execute-advisory] [--advisory-recover-max-rounds <n>] [--advisory-controller-max-cycles <n>] [--plan-only] [--dry-run] [--json]`: governance round-loop runner.
+  - Orchestrates repeated `governance maintain` rounds until target risk or stop condition is reached.
+  - Governance close-loop sessions persist by default at `.kiro/auto/governance-close-loop-sessions`; use `--governance-resume` to continue interrupted governance loops.
+  - Resume defaults inherit persisted governance policy (`target_risk`, `execute_advisory`, `advisory_policy`) to avoid accidental configuration drift; explicit drift is blocked unless `--governance-resume-allow-drift` is set.
+  - `--governance-session-keep` enables automatic post-run governance session pruning (optional age window via `--governance-session-older-than-days`) while preserving the current run session file.
+  - `--execute-advisory` enables automatic advisory action execution when detected (`recover-latest`, `controller-resume-latest`), with autonomous source selection (latest recoverable summary / latest pending controller session) and `skipped` telemetry when no actionable source exists.
+  - Emits round history (`risk_before`/`risk_after`, planned/applicable/applied/failed actions), advisory telemetry (`advisory_*` fields), plus initial/final assessments for auditable autonomous governance convergence.
+- `kse auto governance session list [--limit <n>] [--status <csv>] [--resume-only] [--json]`: inspect persisted governance close-loop session archive (optional resumed-chain filter).
+- `kse auto governance session stats [--days <n>] [--status <csv>] [--resume-only] [--json]`: aggregate governance session completion/failure/convergence/risk telemetry with resumed-session ratio/source breakdown.
+- `kse auto governance session prune [--keep <n>] [--older-than-days <n>] [--dry-run] [--json]`: enforce governance session retention policy.
 
 Recovery memory maintenance:
 - Default recovery memory file: `.kiro/auto/close-loop-recovery-memory.json`

package/docs/command-reference.md

@@ -425,6 +425,10 @@ kse auto session list
 kse auto session list --status completed,partial-failed
 kse auto session list --limit 50 --json
 
+# Aggregate close-loop session telemetry
+kse auto session stats
+kse auto session stats --days 14 --status completed --json
+
 # Prune old close-loop sessions
 kse auto session prune --keep 50
 kse auto session prune --keep 20 --older-than-days 14 --dry-run
@@ -443,6 +447,10 @@ kse auto batch-session list
 kse auto batch-session list --status failed
 kse auto batch-session list --limit 50 --json
 
+# Aggregate close-loop-batch summary telemetry
+kse auto batch-session stats
+kse auto batch-session stats --days 14 --status failed --json
+
 # Prune old close-loop-batch summary sessions
 kse auto batch-session prune --keep 50
 kse auto batch-session prune --keep 20 --older-than-days 14 --dry-run
@@ -452,10 +460,31 @@ kse auto controller-session list
 kse auto controller-session list --status partial-failed
 kse auto controller-session list --limit 50 --json
 
+# Aggregate close-loop-controller summary session telemetry
+kse auto controller-session stats
+kse auto controller-session stats --days 14 --status partial-failed --json
+
 # Prune old close-loop-controller summary sessions
 kse auto controller-session prune --keep 50
 kse auto controller-session prune --keep 20 --older-than-days 14 --dry-run
 
+# Aggregate cross-archive autonomous governance telemetry
+kse auto governance stats
+kse auto governance stats --days 14 --status completed,partial-failed --json
+kse auto governance maintain --session-keep 50 --batch-session-keep 50 --controller-session-keep 50 --json
+kse auto governance maintain --apply --session-keep 20 --batch-session-keep 20 --controller-session-keep 20 --recovery-memory-older-than-days 90 --json
+kse auto governance close-loop --plan-only --max-rounds 3 --target-risk low --json
+kse auto governance close-loop --max-rounds 3 --target-risk low --session-keep 20 --batch-session-keep 20 --controller-session-keep 20 --json
+kse auto governance close-loop --max-rounds 3 --governance-session-keep 50 --governance-session-older-than-days 30 --json
+kse auto governance close-loop --max-rounds 3 --target-risk low --execute-advisory --advisory-recover-max-rounds 3 --advisory-controller-max-cycles 20 --dry-run --json
+kse auto governance close-loop --governance-resume latest --max-rounds 5 --json
+kse auto governance close-loop --governance-resume latest --target-risk high --governance-resume-allow-drift --json
+kse auto governance session list --limit 20 --status completed,failed --json
+kse auto governance session list --resume-only --json
+kse auto governance session stats --days 30 --json
+kse auto governance session stats --resume-only --json
+kse auto governance session prune --keep 50 --older-than-days 30 --dry-run --json
+
 # Recovery memory maintenance
 kse auto recovery-memory show --json
 kse auto recovery-memory scopes --json
@@ -465,11 +494,29 @@ kse auto recovery-memory clear --json
 # Autonomous KPI trend (weekly/daily buckets + CSV export)
 kse auto kpi trend --weeks 12 --period week --mode all --json
 kse auto kpi trend --weeks 8 --period day --mode program --csv --out ./auto-kpi-trend.csv
+
+# Unified observability snapshot (sessions + governance + KPI trend)
+kse auto observability snapshot --days 14 --status completed,failed --json
+kse auto observability snapshot --out .kiro/reports/auto-observability.json --json
+
+# Agent-facing spec interfaces
+kse auto spec status 121-00-master --json
+kse auto spec instructions 121-02-sub-track --json
+
+# Autonomous archive schema compatibility
+kse auto schema check --json
+kse auto schema migrate --json                           # dry-run by default
+kse auto schema migrate --apply --json                  # apply schema_version migration
+kse auto schema migrate --only close-loop-session,batch-session --apply --json
 ``` 
 
 DoD-related options:
 - `--dod-tests <command>`: run a final shell command as a completion gate
 - `--dod-tests-timeout <ms>`: timeout for `--dod-tests` (default `600000`)
+- `--dod-max-risk-level <low|medium|high>`: fail DoD when derived run risk is above threshold
+- `--dod-kpi-min-completion-rate <n>`: minimum close-loop completion rate percent (`0-100`)
+- `--dod-max-success-rate-drop <n>`: max allowed completion-rate drop vs recent baseline (`0-100`)
+- `--dod-baseline-window <n>`: number of recent sessions used for baseline comparison (`1-50`, default `5`)
 - `--dod-tasks-closed`: require no unchecked `- [ ]` items in generated `tasks.md`
 - `--no-dod-docs`: skip doc completeness gate
 - `--no-dod-collab`: skip collaboration completion gate
@@ -486,6 +533,8 @@ DoD-related options:
 - `--replan-attempts <n>`: max automatic replan cycles after failed orchestration (`0-5`, default `1`)
 - `--replan-no-progress-window <n>`: stop replan when no progress repeats for `n` failed cycles (`1-10`, default `3`)
 - `--no-replan`: disable automatic replan cycle
+- `--no-conflict-governance`: disable lease-conflict prediction and scheduling guard
+- `--no-ontology-guidance`: disable scene ontology `agent_hints` scheduling guidance
 
 Close-loop batch (`kse auto close-loop-batch <goals-file>`) options:
 - supports shared close-loop execution options (for example: `--subs`, `--max-parallel`, `--dod*`, `--replan*`, `--dry-run`, `--json`)
@@ -621,8 +670,10 @@ Close-loop recovery (`kse auto close-loop-recover [summary]`) options:
 
 Close-loop session maintenance:
 - `kse auto session list [--limit <n>] [--status <csv>] [--json]`: list persisted close-loop sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto session stats [--days <n>] [--status <csv>] [--json]`: aggregate persisted close-loop session telemetry within an optional recent-day window
 - `kse auto session prune [--keep <n>] [--older-than-days <n>] [--dry-run] [--json]`: prune old session snapshots
-  - JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - List JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - Stats JSON output includes `criteria`, completion/failure rates, `sub_spec_count_sum`, `master_spec_counts`, and `latest_sessions`.
 
 Spec directory maintenance:
 - `kse auto spec-session list [--limit <n>] [--json]`: list persisted spec directories under `.kiro/specs`
@@ -633,13 +684,34 @@ Spec directory maintenance:
 
 Close-loop batch session maintenance:
 - `kse auto batch-session list [--limit <n>] [--status <csv>] [--json]`: list persisted close-loop-batch summary sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto batch-session stats [--days <n>] [--status <csv>] [--json]`: aggregate persisted close-loop-batch summary telemetry within an optional recent-day window
 - `kse auto batch-session prune [--keep <n>] [--older-than-days <n>] [--dry-run] [--json]`: prune old persisted batch summaries
-  - JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - List JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - Stats JSON output includes `criteria`, completion/failure rates, goal-volume sums, processed ratio, and `latest_sessions`.
 
 Close-loop controller session maintenance:
 - `kse auto controller-session list [--limit <n>] [--status <csv>] [--json]`: list persisted close-loop-controller summary sessions (`--status` supports comma-separated, case-insensitive filters)
+- `kse auto controller-session stats [--days <n>] [--status <csv>] [--json]`: aggregate persisted close-loop-controller status/throughput telemetry within an optional recent-day window
 - `kse auto controller-session prune [--keep <n>] [--older-than-days <n>] [--dry-run] [--json]`: prune old persisted controller summaries
-  - JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - List JSON output includes `status_filter` and `status_counts` over filtered sessions.
+  - Stats JSON output includes `criteria`, `status_counts`, `queue_format_counts`, completion/failure rates, goal-volume sums, and `latest_sessions`.
+
+Cross-archive autonomous governance maintenance:
+- `kse auto governance stats [--days <n>] [--status <csv>] [--json]`: aggregate a unified governance snapshot from session/batch-session/controller-session archives plus recovery memory state.
+  - JSON output includes `totals`, `throughput`, `health` (`risk_level`, `concerns`, `recommendations`), `top_master_specs`, `recovery_memory`, and full per-archive stats under `archives`.
+- `kse auto governance maintain [--days <n>] [--status <csv>] [--session-keep <n>] [--batch-session-keep <n>] [--controller-session-keep <n>] [--recovery-memory-older-than-days <n>] [--apply] [--dry-run] [--json]`: run governance-maintenance planning and optional execution in one command.
+  - Plan-only mode is default; add `--apply` to execute maintenance actions (`session prune`, `batch-session prune`, `controller-session prune`, `recovery-memory prune`).
+  - JSON output includes `assessment` (pre-maintenance governance snapshot), `plan`, `executed_actions`, `summary`, and `after_assessment` (only when `--apply` without `--dry-run`).
+- `kse auto governance close-loop [--days <n>] [--status <csv>] [--session-keep <n>] [--batch-session-keep <n>] [--controller-session-keep <n>] [--recovery-memory-older-than-days <n>] [--max-rounds <n>] [--target-risk <low|medium|high>] [--governance-resume <session|latest|file>] [--governance-resume-allow-drift] [--governance-session-id <id>] [--no-governance-session] [--governance-session-keep <n>] [--governance-session-older-than-days <n>] [--execute-advisory] [--advisory-recover-max-rounds <n>] [--advisory-controller-max-cycles <n>] [--plan-only] [--dry-run] [--json]`: run governance rounds until stop condition (target risk reached, no actionable maintenance/advisory, non-mutating mode, maintenance/advisory failures, or max rounds).
+  - `--plan-only` runs a single non-mutating planning round.
+  - Governance close-loop sessions are persisted by default at `.kiro/auto/governance-close-loop-sessions/*.json`; use `--governance-resume` to continue interrupted governance loops.
+  - On resume, KSE reuses persisted policy defaults (`target_risk`, `execute_advisory`, `advisory_policy`) unless explicitly overridden. Explicit policy drift is blocked by default; add `--governance-resume-allow-drift` to force override.
+  - `--governance-session-keep` (with optional `--governance-session-older-than-days`) enables post-run governance session retention pruning while protecting the current session snapshot.
+  - `--execute-advisory` enables automatic advisory action execution (`recover-latest`, `controller-resume-latest`) when governance assessment detects failed sessions or controller pending goals; KSE auto-selects the latest actionable advisory source and reports `skipped` (not `failed`) when no actionable source exists.
+  - JSON output includes round-by-round risk/action telemetry (`rounds`), advisory telemetry (`execute_advisory`, `advisory_policy`, `advisory_summary`, `rounds[*].advisory_actions`), plus `initial_assessment`, `final_assessment`, and convergence metadata.
+- `kse auto governance session list [--limit <n>] [--status <csv>] [--resume-only] [--json]`: list persisted governance close-loop sessions (`--resume-only` filters to resumed-chain sessions only).
+- `kse auto governance session stats [--days <n>] [--status <csv>] [--resume-only] [--json]`: aggregate governance close-loop session telemetry (completion/failure/convergence, rounds, risk/stop composition, and resumed-chain ratios/source counts).
+- `kse auto governance session prune [--keep <n>] [--older-than-days <n>] [--dry-run] [--json]`: prune governance close-loop session archive by retention policy.
 
 Close-loop recovery memory maintenance:
 - `kse auto recovery-memory show [--scope <scope>] [--json]`: inspect persisted recovery signatures/actions and aggregate stats (optionally scoped)
@@ -653,6 +725,19 @@ Autonomous KPI trend:
   - `--csv` prints CSV rows to stdout and writes CSV when used with `--out` (JSON remains default).
   - JSON output includes `mode_breakdown` (batch/program/recover/controller/other run distribution), `anomaly_detection`, and flattened `anomalies` (latest-period regression checks against historical baseline).
 
+Unified observability snapshot:
+- `kse auto observability snapshot [--days <n>] [--status <csv>] [--weeks <n>] [--trend-mode <mode>] [--trend-period <period>] [--out <path>] [--json]`: generate one unified observability snapshot that combines close-loop session stats, batch stats, controller stats, governance session stats, governance health, and KPI trend.
+- JSON output includes top-level `highlights` plus detailed archive/trend payloads under `snapshots`.
+
+Agent-facing spec interfaces:
+- `kse auto spec status <spec-name> [--json]`: structured status for one spec (`docs`, `task_progress`, `collaboration`, `health`).
+- `kse auto spec instructions <spec-name> [--json]`: machine-readable execution instructions for one spec (`next_actions`, `priority_open_tasks`, recommended commands, document excerpts).
+
+Autonomous archive schema compatibility:
+- `kse auto schema check [--only <scopes>] [--json]`: scan archive schema compatibility (`schema_version`) for `close-loop-session`, `batch-session`, `controller-session`, and `governance-session`.
+- `kse auto schema migrate [--only <scopes>] [--target-version <version>] [--apply] [--json]`: migrate/backfill `schema_version` across autonomous archives.
+  - Default mode is dry-run; use `--apply` to persist changes.
+
 Recommended `.kiro/config/orchestrator.json`:
 
 ```json
@@ -731,6 +816,14 @@ kse scene ontology show --package ./my-scene-package --json
 kse scene ontology deps --package <path> --ref <node-ref>
 kse scene ontology deps --package ./my-scene-package --ref entity:Order --json
 
+# Analyze reverse dependency impact radius (what will be affected)
+kse scene ontology impact --package <path> --ref <node-ref>
+kse scene ontology impact --package ./my-scene-package --ref service:createOrder --relation depends_on,composes --max-depth 2 --json
+
+# Find shortest ontology relation path between two refs
+kse scene ontology path --package <path> --from <source-ref> --to <target-ref>
+kse scene ontology path --package ./my-scene-package --from service:createOrder --to entity:Order --undirected --json
+
 # Validate ontology graph (detect dangling edges, cycles)
 kse scene ontology validate --package <path>
 kse scene ontology validate --package ./my-scene-package --json

package/docs/scene-runtime-guide.md

@@ -268,6 +268,12 @@ Defined in `governance_contract.data_lineage`:
 }
 ```
 
+### Impact & Path Queries
+
+Use ontology as an operational graph, not just static visualization:
+- `impact`: reverse-traverse matching relations to estimate blast radius from one changed ref
+- `path`: find the shortest relation path between two refs for dependency explainability
+
 ### Agent Hints
 
 The `agent_hints` field provides metadata for autonomous AI operation:
@@ -302,6 +308,14 @@ kse scene ontology show --package ./my-package --json
 kse scene ontology deps --package <path> --ref <node-ref>
 kse scene ontology deps --package ./my-package --ref entity:Order --json
 
+# Query reverse impact radius
+kse scene ontology impact --package <path> --ref <node-ref>
+kse scene ontology impact --package ./my-package --ref service:createOrder --relation depends_on,composes --max-depth 2 --json
+
+# Query shortest relation path between two refs
+kse scene ontology path --package <path> --from <source-ref> --to <target-ref>
+kse scene ontology path --package ./my-package --from service:createOrder --to entity:Order --undirected --json
+
 # Validate graph (dangling edges, cycles)
 kse scene ontology validate --package <path>
 kse scene ontology validate --package ./my-package --json