snipara-companion 3.6.13 → 3.6.14

package/CHANGELOG.md

@@ -2,6 +2,20 @@
 
 Release notes for `snipara-companion`, newest first.
 
+## Unreleased
+
+## New In 3.6.14
+
+- Adds opt-in `--minimum-change-mode review` forwarding for hosted context
+  queries and code-impact reviews, including the evidence-backed
+  minimum-safe-change receipt in answer packs.
+
+- Lets managed `pre-commit` and `pre-push` hooks receive an explicit,
+  one-operation review-only acknowledgement through
+  `SNIPARA_COLLABORATION_GUARD_ACK_REVIEW_ONLY=1`; the hook immediately
+  re-runs without the flag so blocking leases and hard conflicts remain
+  enforced.
+
 ## New In 3.6.13
 
 - Adds a native spec-driven `feature` workflow for durable `spec.md`, `plan.md`,

package/README.md

@@ -18,7 +18,7 @@ state, handoffs, and durable task outcomes.
 ```bash
 npx -y create-snipara@latest init --client cursor --starter
 npx -y snipara-companion session-bootstrap --include-session-context --max-context-tokens 1000
-npx -y snipara-companion impact src/auth/session.ts --source local
+npx -y snipara-companion code impact --changed-files src/auth/session.ts --source local
 npx -y snipara-companion source init .
 ```
 
@@ -106,7 +106,7 @@ These commands are useful without hosted Snipara:
 | Command                                                                               | What it gives you locally                                                    |
 | ------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------- |
 | `source init` / `source sync` / `source status`                                       | Local source snapshot, document preview, and code overlay                    |
-| `impact` / `code impact`                                                              | Bounded transitive blast radius, chains, and explainable local risk          |
+| `code impact` (`impact` alias) / `code local impact`                                  | Bounded transitive blast radius, chains, and explainable local risk          |
 | `reality-check`                                                                       | Intent Ledger, Unknown Registry, auto-linked context, and inspectable proof  |
 | `code callers` / `imports` / `neighbors` / `shortest-path`                            | Structural repo questions from local files                                   |
 | `workflow start` / `phase-start` / `phase-commit` / `resume`                          | Agent continuity that survives compaction                                    |
@@ -125,6 +125,32 @@ These commands are useful without hosted Snipara:
 | `intelligence ledger-export`                                                          | Structured redacted ledger JSON for replay and review                        |
 | `stuck-guard`, `memory-guard`, `pre-tool`, `post-tool`                                | Fail-soft local guards and hook helpers                                      |
 
+### Canonical command forms
+
+Use the exact CLI value after `workflow run --mode`:
+
+| CLI value     | Guide label         | Meaning                                                               |
+| ------------- | ------------------- | --------------------------------------------------------------------- |
+| `lite`        | LITE                | Small, known-scope work with no mandatory hosted context call         |
+| `standard`    | STANDARD            | Normal work with context and code-graph follow-up when needed         |
+| `auto`        | AUTO                | Routes by task intent to `lite`, `standard`, `full`, or `orchestrate` |
+| `full`        | FULL                | Managed, phased work with durable context and plan support            |
+| `orchestrate` | FULL + ORCHESTRATED | Explicit deeper orchestration for multi-agent or proof-gate work      |
+
+The root `run` command is the Project Intelligence judgment/release flow.
+`workflow run` is the workflow-mode runner. They share a verb but are not
+aliases.
+
+Use `code impact` as the canonical impact gate. The root `impact` command is a
+compatibility alias, while `code local impact` is the separate non-canonical
+local-overlay query. Use `final-commit` as the canonical final workflow
+closeout; `workflow final-commit` remains a compatibility alias. `task-commit`
+captures a durable task outcome, and `workflow phase-commit` records one
+managed phase and advances the workflow; these are distinct operations.
+
+For a public, copy/paste-oriented table, see the
+[Companion CLI Command Reference](https://www.snipara.com/docs/integration/companion/reference).
+
 ### Spec-driven feature workflow
 
 Companion provides a native Spec Kit-style preparation layer without creating a
@@ -506,14 +532,14 @@ or ahead work; `--source local` remains completely offline.
 Use `npx` for one-off checks:
 
 ```bash
-npx -y snipara-companion impact src/auth/session.ts --source local
+npx -y snipara-companion code impact --changed-files src/auth/session.ts --source local
 ```
 
 Install globally only if you use it every day:
 
 ```bash
 npm install -g snipara-companion
-snipara-companion impact src/auth/session.ts
+snipara-companion code impact --changed-files src/auth/session.ts
 snipara-companion workflow resume
 ```
 

package/dist/index.d.ts

@@ -89,6 +89,7 @@ interface ContextQueryResult {
     };
 }
 type ContextQuerySearchMode = "keyword" | "semantic" | "hybrid";
+type MinimumChangeMode = "off" | "review";
 interface ContextQueryOptions {
     searchMode?: ContextQuerySearchMode;
     includeMetadata?: boolean;
@@ -97,6 +98,7 @@ interface ContextQueryOptions {
     includeSharedContext?: boolean;
     includeAllTiers?: boolean;
     returnReferences?: boolean;
+    minimumChangeMode?: MinimumChangeMode;
 }
 interface AnswerPackSource {
     title?: string;
@@ -125,6 +127,7 @@ interface AnswerPack {
         graph_context_included?: boolean;
         impact_hint?: string | null;
     };
+    minimum_safe_change?: Record<string, unknown>;
     token_count?: number;
 }
 interface HostedContextControlDiffResponse {
@@ -1141,6 +1144,7 @@ declare class RLMClient {
         direction?: "in" | "out" | "both";
         edgeKinds?: string[];
         limit?: number;
+        minimumChangeMode?: MinimumChangeMode;
     }): Promise<Record<string, unknown>>;
     /**
      * Track accessed files using snipara_remember
@@ -2198,6 +2202,7 @@ declare function writeLocalCodeOverlayCache(manifest: LocalCodeOverlayManifest):
 declare function readLocalCodeOverlayCache(cwd?: string): LocalCodeOverlayManifest | null;
 declare function writeLocalCodePromotionState(state: LocalCodePromotionState): string;
 declare function readLocalCodePromotionState(cwd?: string): LocalCodePromotionState | null;
+declare function resolveMinimumChangeMode(value: unknown): MinimumChangeMode;
 declare function resolveCodeGraphMode(args: {
     requested: CodeGraphSource;
     dirtyFiles: string[];
@@ -5373,4 +5378,4 @@ declare function installAutomationBundle(args: {
 }): Promise<AutomationInstallResult>;
 declare function getAutomationStatus(projectDir?: string): AutomationStatusResult;
 
-export { AGENT_CONTEXT_EVIDENCE_DEFAULT_LEDGER, AUTOMATION_MANIFEST_RELATIVE_PATH, AutomationInstallConflictError, AutomationUnsupportedHookBundleError, CODING_INTELLIGENCE_LEDGER_VERSION, COLLABORATION_STATE_RELATIVE_PATH, ENGINEERING_LEAD_EXECUTION_RECEIPT_STAGES, ENGINEERING_LEAD_EXECUTION_RECEIPT_STATUSES, ENGINEERING_LEAD_POSTURES, ENGINEERING_LEAD_PROOF_VERIFICATION_SOURCES, ENGINEERING_LEAD_PROOF_VERIFICATION_STATUSES, ENGINEERING_LEAD_ROUTING_MODES, ENGINEERING_LEAD_STATUSES, ENGINEERING_LEAD_SUPERVISION_STATUSES, ENGINEERING_LEAD_WORKER_ROLES, ENGINEERING_LEAD_WORK_PACKAGE_STATUSES, FEATURE_DEFAULT_OUTPUT_DIR, FEATURE_WORK_ITEM_VERSION, FINAL_COMMIT_REPORT_RELATIVE_PATH, FINAL_COMMIT_REPORT_VERSION, MEMORY_GUARD_CONTEXT_TIMEOUT_MS, MEMORY_GUARD_RECALL_TIMEOUT_MS, ORCHESTRATOR_HANDOFF_RELATIVE_PATH, POLICY_LEDGER_SYNC_REPORT_VERSION, PRODUCER_LOOP_ARTIFACT_VERSION, PRODUCER_LOOP_RELATIVE_DIR, PRODUCER_LOOP_REPORT_VERSION, TEAM_SYNC_STATE_RELATIVE_PATH, WHY_OUTCOME_CAPTURE_VERSION, WORKFLOW_PLANS_RELATIVE_DIR, WORKFLOW_STATE_RELATIVE_PATH, addLocalWorker, agentContextEvidenceCollectCommand, agentContextEvidenceRecordCommand, agentContextEvidenceStatusCommand, agentContextEvidenceTemplateCommand, agentContextResolveCommand, agentContextValidateCommand, agentReadinessAuditCommand, appendActivityEvent, applyLocalContextMutationPlan, applyManagedWorkflowPolicyAutoResponses, archiveInactiveTeamSyncWork, attachLocalContextPackReceipts, autoArchiveTeamSyncState, buildAdaptiveWorkRoutingRecommendation, buildAgentReadinessAuditReport, buildAgenticHandoffMarkdown, buildAgenticTimeline, buildAgenticWorkStatus, buildAutomationInstallPlan, buildCanonicalEvent, buildCodeHooksInstallPlan, buildCodePromotionResult, buildCodeStatusResult, buildCodeSyncResult, buildCodingIntelligenceLedger, buildCollaborationActor, buildCollaborationGuardActionCards, buildCollaborationHooksInstallPlan, buildCommitResultMetadata, buildCompanionEngineeringLeadPlanReport, buildContextPackStats, buildEvalCaseArtifact, buildFeaturePlanMarkdown, buildFeatureSpecMarkdown, buildFeatureTasksMarkdown, buildFinalCommitReport, buildGeneratedWorkflowPlanDocument, buildHostedCodeOverlayUploadPayload, buildHostedGuardPayload, buildJournalCheckpointEntry, buildLocalAgentContextValidationReport, buildLocalCallersResult, buildLocalCodeOverlay, buildLocalContextMutationPlan, buildLocalContextPackReceipt, buildLocalContextPackReceipts, buildLocalImpactResult, buildLocalImportsResult, buildLocalNeighborsResult, buildLocalProjectContextValidationReport, buildLocalProjectDriftReport, buildLocalProjectRealityCheck, buildLocalProjectRealityCheckWithAutoContext, buildLocalShortestPathResult, buildLocalSourceSnapshot, buildLocalSourceStatus, buildLocalSourceSyncResult, buildMemoryAudit, buildOnboardFolderManifest, buildOrchestratorHandoff, buildPolicyLedgerSyncReport, buildProducerLoopReport, buildProjectIntelligenceBrief, buildProjectIntelligenceRun, buildProjectJudgmentCard, buildSessionBootstrapBrief, buildSessionBootstrapQuality, buildSessionSnapshot, buildSyncDocumentsDryRun, buildTeamSyncHandoffRecord, buildTeamSyncStartWorkRecord, buildTeamSyncSummary, buildToolCallPayload, buildToolResultPayload, buildVerificationPlan, buildWhyOutcomeCaptureReport, buildWorkerTrustCandidates, buildWorkflowImpactGate, buildWorkflowPhaseCommitReceipt, buildWorkflowPhaseCommitSummary, buildWorkflowPlanScaffold, captureStandaloneCommitWhy, categoryFromGuardTag, classifyToolResult, cleanContextPacks, codingLedgerExportCommand, collaborationIdeStatusCommand, collectAgentReadinessLocalSignals, collectPolicyLedgerSyncArtifacts, collectSyncDocuments, collectSyncDocumentsInput, compactHostedGuardResources, compareLocalSourceSnapshots, completeTeamSyncStateFromEvidence, completeTeamSyncWorkFromEvidence, contextControlApplyCommand, contextControlDriftCommand, contextControlHostedApplyCommand, contextControlHostedDiffCommand, contextControlPlanCommand, contextControlValidateCommand, contextPackCleanCommand, contextPackPackCommand, contextPackRetrieveCommand, contextPackStatsCommand, controlledWorkerExecuteCommand, createClient, createEmptyCollaborationState, createEmptyTeamSyncState, createLocalQueryCache, deriveLocalCollaborationResourcesFromFiles, deriveManagedWorkflowJudgmentResolution, detectReleaseSurfacesFromFiles, detectRuntimeEnvironment, evalExportCommand, evalRunCommand, evaluateProjectPolicyGates, extractCommandFromToolInput, extractFilesFromToolInput, featureInitCommand, featurePlanCommand, featureSpecifyCommand, featureStartCommand, featureStatusCommand, featureTasksCommand, formatAgentContextResolution, formatAgentReadinessAuditReport, formatCompanionEngineeringLeadPlanReport, formatFinalCommitReport, formatOrchestratorRecommendationReason, formatPolicyGateDecision, formatProjectJudgmentCard, formatStuckGuardDecision, getAutomationManifestPath, getAutomationStatus, getCollaborationStatePath, getConfigPath, getContextPackStoragePaths, getLocalCodeOverlayCachePath, getLocalCodePromotionStatePath, getLocalSourceSnapshotPath, getOrchestratorRecommendation, getPlanStepDisplayTitle, getStagedFiles, getStuckGuardInjection, getTeamSyncStatePath, hasActiveManagedWorkflow, hashWorkerProfile, ingestReferences, installAutomationBundle, leadPlanCommand, listProjectsForApiKey, loadAutomationManifest, loadCollaborationState, loadConfig, loadTeamSyncState, memoryAuditCommand, memoryCleanCandidatesCommand, memoryCompactCommand, memoryHealthCommand, memoryLocalCommand, memoryReviewsCommand, mergeHybridCodeResults, normalizeCollaborationFiles, normalizeFeatureSlug, normalizeGuardTag, normalizeWorkflowPlanInput, outcomeCapturePreviewCommand, packContext, parseCollaborationResources, projectIntelligenceBriefCommand, projectIntelligenceRunCommand, readActivityTimeline, readAgentContextEvidenceLedger, readLocalCodeOverlayCache, readLocalCodePromotionState, readLocalSourceSnapshot, readLocalWorkersConfig, readSessionSnapshot, readStandaloneCommitEvidence, readWorkerTrustEvent, realityCheckCommand, referencesIngestCommand, referencesScanCommand, resolveAutoWorkflowMode, resolveCodeGraphMode, resolveContextPackRecord, resolveFeatureArtifactPaths, resolveFullWorkflowTokenBudget, resolveLocalWorkerRoutingDefaults, resolveQueryFromToolInput, retrieveContextPack, runMemoryGuardCheck, saveCollaborationState, saveConfig, saveTeamSyncState, scanReferences, shouldFailCollaborationGuard, shouldSuggestOrchestratorForWorkflow, shouldSuggestRuntimeForWorkflow, summarizeLocalCodeOverlay, teamSyncSweepCommand, validatePlanResult, verifyCommand, workerTrustCandidateCommand, workerTrustReviewCommand, workerTrustStatusCommand, workersLocalAddCommand, workersLocalListCommand, workersLocalProbePrintCommand, workersLocalRemoveCommand, workersLocalStatusCommand, workflowSyncPolicyLedgerCommand, writeFinalCommitReport, writeLocalCodeOverlayCache, writeLocalCodePromotionState, writeLocalSourceSnapshot, writeOrchestratorHandoff, writeProducerLoopArtifact, writeSessionSnapshot };
+export { AGENT_CONTEXT_EVIDENCE_DEFAULT_LEDGER, AUTOMATION_MANIFEST_RELATIVE_PATH, AutomationInstallConflictError, AutomationUnsupportedHookBundleError, CODING_INTELLIGENCE_LEDGER_VERSION, COLLABORATION_STATE_RELATIVE_PATH, ENGINEERING_LEAD_EXECUTION_RECEIPT_STAGES, ENGINEERING_LEAD_EXECUTION_RECEIPT_STATUSES, ENGINEERING_LEAD_POSTURES, ENGINEERING_LEAD_PROOF_VERIFICATION_SOURCES, ENGINEERING_LEAD_PROOF_VERIFICATION_STATUSES, ENGINEERING_LEAD_ROUTING_MODES, ENGINEERING_LEAD_STATUSES, ENGINEERING_LEAD_SUPERVISION_STATUSES, ENGINEERING_LEAD_WORKER_ROLES, ENGINEERING_LEAD_WORK_PACKAGE_STATUSES, FEATURE_DEFAULT_OUTPUT_DIR, FEATURE_WORK_ITEM_VERSION, FINAL_COMMIT_REPORT_RELATIVE_PATH, FINAL_COMMIT_REPORT_VERSION, MEMORY_GUARD_CONTEXT_TIMEOUT_MS, MEMORY_GUARD_RECALL_TIMEOUT_MS, ORCHESTRATOR_HANDOFF_RELATIVE_PATH, POLICY_LEDGER_SYNC_REPORT_VERSION, PRODUCER_LOOP_ARTIFACT_VERSION, PRODUCER_LOOP_RELATIVE_DIR, PRODUCER_LOOP_REPORT_VERSION, TEAM_SYNC_STATE_RELATIVE_PATH, WHY_OUTCOME_CAPTURE_VERSION, WORKFLOW_PLANS_RELATIVE_DIR, WORKFLOW_STATE_RELATIVE_PATH, addLocalWorker, agentContextEvidenceCollectCommand, agentContextEvidenceRecordCommand, agentContextEvidenceStatusCommand, agentContextEvidenceTemplateCommand, agentContextResolveCommand, agentContextValidateCommand, agentReadinessAuditCommand, appendActivityEvent, applyLocalContextMutationPlan, applyManagedWorkflowPolicyAutoResponses, archiveInactiveTeamSyncWork, attachLocalContextPackReceipts, autoArchiveTeamSyncState, buildAdaptiveWorkRoutingRecommendation, buildAgentReadinessAuditReport, buildAgenticHandoffMarkdown, buildAgenticTimeline, buildAgenticWorkStatus, buildAutomationInstallPlan, buildCanonicalEvent, buildCodeHooksInstallPlan, buildCodePromotionResult, buildCodeStatusResult, buildCodeSyncResult, buildCodingIntelligenceLedger, buildCollaborationActor, buildCollaborationGuardActionCards, buildCollaborationHooksInstallPlan, buildCommitResultMetadata, buildCompanionEngineeringLeadPlanReport, buildContextPackStats, buildEvalCaseArtifact, buildFeaturePlanMarkdown, buildFeatureSpecMarkdown, buildFeatureTasksMarkdown, buildFinalCommitReport, buildGeneratedWorkflowPlanDocument, buildHostedCodeOverlayUploadPayload, buildHostedGuardPayload, buildJournalCheckpointEntry, buildLocalAgentContextValidationReport, buildLocalCallersResult, buildLocalCodeOverlay, buildLocalContextMutationPlan, buildLocalContextPackReceipt, buildLocalContextPackReceipts, buildLocalImpactResult, buildLocalImportsResult, buildLocalNeighborsResult, buildLocalProjectContextValidationReport, buildLocalProjectDriftReport, buildLocalProjectRealityCheck, buildLocalProjectRealityCheckWithAutoContext, buildLocalShortestPathResult, buildLocalSourceSnapshot, buildLocalSourceStatus, buildLocalSourceSyncResult, buildMemoryAudit, buildOnboardFolderManifest, buildOrchestratorHandoff, buildPolicyLedgerSyncReport, buildProducerLoopReport, buildProjectIntelligenceBrief, buildProjectIntelligenceRun, buildProjectJudgmentCard, buildSessionBootstrapBrief, buildSessionBootstrapQuality, buildSessionSnapshot, buildSyncDocumentsDryRun, buildTeamSyncHandoffRecord, buildTeamSyncStartWorkRecord, buildTeamSyncSummary, buildToolCallPayload, buildToolResultPayload, buildVerificationPlan, buildWhyOutcomeCaptureReport, buildWorkerTrustCandidates, buildWorkflowImpactGate, buildWorkflowPhaseCommitReceipt, buildWorkflowPhaseCommitSummary, buildWorkflowPlanScaffold, captureStandaloneCommitWhy, categoryFromGuardTag, classifyToolResult, cleanContextPacks, codingLedgerExportCommand, collaborationIdeStatusCommand, collectAgentReadinessLocalSignals, collectPolicyLedgerSyncArtifacts, collectSyncDocuments, collectSyncDocumentsInput, compactHostedGuardResources, compareLocalSourceSnapshots, completeTeamSyncStateFromEvidence, completeTeamSyncWorkFromEvidence, contextControlApplyCommand, contextControlDriftCommand, contextControlHostedApplyCommand, contextControlHostedDiffCommand, contextControlPlanCommand, contextControlValidateCommand, contextPackCleanCommand, contextPackPackCommand, contextPackRetrieveCommand, contextPackStatsCommand, controlledWorkerExecuteCommand, createClient, createEmptyCollaborationState, createEmptyTeamSyncState, createLocalQueryCache, deriveLocalCollaborationResourcesFromFiles, deriveManagedWorkflowJudgmentResolution, detectReleaseSurfacesFromFiles, detectRuntimeEnvironment, evalExportCommand, evalRunCommand, evaluateProjectPolicyGates, extractCommandFromToolInput, extractFilesFromToolInput, featureInitCommand, featurePlanCommand, featureSpecifyCommand, featureStartCommand, featureStatusCommand, featureTasksCommand, formatAgentContextResolution, formatAgentReadinessAuditReport, formatCompanionEngineeringLeadPlanReport, formatFinalCommitReport, formatOrchestratorRecommendationReason, formatPolicyGateDecision, formatProjectJudgmentCard, formatStuckGuardDecision, getAutomationManifestPath, getAutomationStatus, getCollaborationStatePath, getConfigPath, getContextPackStoragePaths, getLocalCodeOverlayCachePath, getLocalCodePromotionStatePath, getLocalSourceSnapshotPath, getOrchestratorRecommendation, getPlanStepDisplayTitle, getStagedFiles, getStuckGuardInjection, getTeamSyncStatePath, hasActiveManagedWorkflow, hashWorkerProfile, ingestReferences, installAutomationBundle, leadPlanCommand, listProjectsForApiKey, loadAutomationManifest, loadCollaborationState, loadConfig, loadTeamSyncState, memoryAuditCommand, memoryCleanCandidatesCommand, memoryCompactCommand, memoryHealthCommand, memoryLocalCommand, memoryReviewsCommand, mergeHybridCodeResults, normalizeCollaborationFiles, normalizeFeatureSlug, normalizeGuardTag, normalizeWorkflowPlanInput, outcomeCapturePreviewCommand, packContext, parseCollaborationResources, projectIntelligenceBriefCommand, projectIntelligenceRunCommand, readActivityTimeline, readAgentContextEvidenceLedger, readLocalCodeOverlayCache, readLocalCodePromotionState, readLocalSourceSnapshot, readLocalWorkersConfig, readSessionSnapshot, readStandaloneCommitEvidence, readWorkerTrustEvent, realityCheckCommand, referencesIngestCommand, referencesScanCommand, resolveAutoWorkflowMode, resolveCodeGraphMode, resolveContextPackRecord, resolveFeatureArtifactPaths, resolveFullWorkflowTokenBudget, resolveLocalWorkerRoutingDefaults, resolveMinimumChangeMode, resolveQueryFromToolInput, retrieveContextPack, runMemoryGuardCheck, saveCollaborationState, saveConfig, saveTeamSyncState, scanReferences, shouldFailCollaborationGuard, shouldSuggestOrchestratorForWorkflow, shouldSuggestRuntimeForWorkflow, summarizeLocalCodeOverlay, teamSyncSweepCommand, validatePlanResult, verifyCommand, workerTrustCandidateCommand, workerTrustReviewCommand, workerTrustStatusCommand, workersLocalAddCommand, workersLocalListCommand, workersLocalProbePrintCommand, workersLocalRemoveCommand, workersLocalStatusCommand, workflowSyncPolicyLedgerCommand, writeFinalCommitReport, writeLocalCodeOverlayCache, writeLocalCodePromotionState, writeLocalSourceSnapshot, writeOrchestratorHandoff, writeProducerLoopArtifact, writeSessionSnapshot };

package/dist/index.js

@@ -253,6 +253,7 @@ __export(index_exports, {
   resolveFeatureArtifactPaths: () => resolveFeatureArtifactPaths,
   resolveFullWorkflowTokenBudget: () => resolveFullWorkflowTokenBudget,
   resolveLocalWorkerRoutingDefaults: () => resolveLocalWorkerRoutingDefaults,
+  resolveMinimumChangeMode: () => resolveMinimumChangeMode,
   resolveQueryFromToolInput: () => resolveQueryFromToolInput,
   retrieveContextPack: () => retrieveContextPack,
   runMemoryGuardCheck: () => runMemoryGuardCheck,
@@ -975,7 +976,8 @@ var RLMClient = class {
       auto_decompose: options.autoDecompose,
       include_shared_context: options.includeSharedContext,
       include_all_tiers: options.includeAllTiers,
-      return_references: options.returnReferences
+      return_references: options.returnReferences,
+      minimum_change_mode: options.minimumChangeMode ?? "off"
     });
     return {
       sections: (result.sections || []).map((s) => ({
@@ -1072,7 +1074,8 @@ var RLMClient = class {
       depth: options.depth ?? 3,
       direction: options.direction ?? "both",
       edge_kinds: options.edgeKinds,
-      limit: options.limit ?? 50
+      limit: options.limit ?? 50,
+      minimum_change_mode: options.minimumChangeMode ?? "off"
     });
   }
   /**
@@ -5328,6 +5331,11 @@ function resolveRequestedCodeGraphSource(value) {
   }
   throw new Error("--source must be one of: auto, local, hosted, hybrid");
 }
+function resolveMinimumChangeMode(value) {
+  if (value === void 0 || value === null || value === "") return "off";
+  if (value === "off" || value === "review") return value;
+  throw new Error("--minimum-change-mode must be one of: off, review");
+}
 function resolveCodeGraphMode(args) {
   if (args.fallbackHosted && args.requested !== "hosted") {
     return args.hostedConfigured ? { selected: "hybrid_graph", reason: "fallback_hosted_requested" } : { selected: "local_overlay", reason: "fallback_hosted_not_configured" };
@@ -5484,6 +5492,7 @@ function mergeHybridCodeResults(verb, localResult, hostedResult, hostedError) {
   );
   const coverageGaps = mergeResultArrays([hosted, local], ["coverage_gaps", "coverageGaps"]);
   const warnings = mergeResultArrays([hosted, local], ["warnings"]);
+  const minimumSafeChange = hosted.minimum_safe_change ?? hosted.minimumSafeChange ?? local.minimum_safe_change;
   return {
     title: `Hybrid ${verb}`,
     mode: hostedError ? "hybrid_degraded_local" : "hybrid",
@@ -5506,6 +5515,7 @@ function mergeHybridCodeResults(verb, localResult, hostedResult, hostedError) {
     recommended_actions: recommendedActions,
     coverage_gaps: coverageGaps,
     warnings,
+    ...minimumSafeChange ? { minimum_safe_change: minimumSafeChange } : {},
     ...hosted.index_freshness !== void 0 ? { index_freshness: hosted.index_freshness } : {},
     ...hosted.impact !== void 0 ? { impact: hosted.impact } : {},
     local: localResult,
@@ -5587,7 +5597,8 @@ async function callHostedCodeTool(verb, options) {
         depth: options.depth,
         direction: traversalDirection(options.direction),
         edgeKinds: options.edgeKinds,
-        limit: options.limit
+        limit: options.limit,
+        minimumChangeMode: options.minimumChangeMode
       });
     }
   }
@@ -20469,7 +20480,22 @@ function buildCollaborationGitHookBlock(hookName) {
     "    fi",
     "  }",
     `  snipara_companion_hook collaboration claim --resource SURFACE:checkout-git-write --mode EXCLUSIVE --ttl-seconds 300 --reason "${profile} git write" --json >/dev/null || true`,
-    `  snipara_companion_hook collaboration guard --profile ${profile} --action ${profile} --resource SURFACE:checkout-git-write --enforce --json >/dev/null`,
+    "  snipara_companion_guard() {",
+    `    snipara_companion_hook collaboration guard --profile ${profile} --action ${profile} --resource SURFACE:checkout-git-write --enforce "$@" --json >/dev/null`,
+    "  }",
+    '  if [ "${SNIPARA_COLLABORATION_GUARD_ACK_REVIEW_ONLY:-0}" = "1" ]; then',
+    "    snipara_companion_guard --ack-review-only",
+    "    snipara_guard_status=$?",
+    '    if [ "$snipara_guard_status" -eq 0 ]; then',
+    "      snipara_companion_guard",
+    "      snipara_guard_status=$?",
+    "    fi",
+    '    if [ "$snipara_guard_status" -ne 0 ]; then',
+    '      exit "$snipara_guard_status"',
+    "    fi",
+    "  else",
+    "    snipara_companion_guard",
+    "  fi",
     "fi",
     collaborationHookBlockMarker(hookName, "end")
   ].join("\n");
@@ -32606,6 +32632,7 @@ async function queryCommand(options) {
   const result = await client.queryContext(options.query, options.maxTokens || 8e3, {
     searchMode,
     includeAnswerPack: options.includeAnswerPack,
+    minimumChangeMode: options.minimumChangeMode,
     autoDecompose: options.autoDecompose,
     includeSharedContext: options.includeSharedContext
   });
@@ -43943,7 +43970,7 @@ program.command("memory-guard").description(
 );
 program.command("query").description(
   "Search project documents, parsed files, and current truth through hosted Snipara context"
-).requiredOption("-q, --query <query>", "Search query").option("-m, --max-tokens <number>", "Maximum tokens", "8000").option("--search-mode <mode>", "Search mode: keyword|semantic|hybrid", "hybrid").option("--timeout-ms <number>", "Hosted context query timeout in milliseconds", "30000").option("--no-answer-pack", "Skip Answer Pack generation").option("--no-auto-decompose", "Disable automatic query decomposition").option("--no-shared-context", "Exclude linked shared context").option(
+).requiredOption("-q, --query <query>", "Search query").option("-m, --max-tokens <number>", "Maximum tokens", "8000").option("--search-mode <mode>", "Search mode: keyword|semantic|hybrid", "hybrid").option("--timeout-ms <number>", "Hosted context query timeout in milliseconds", "30000").option("--no-answer-pack", "Skip Answer Pack generation").option("--minimum-change-mode <mode>", "off|review (opt-in non-blocking review)", "off").option("--no-auto-decompose", "Disable automatic query decomposition").option("--no-shared-context", "Exclude linked shared context").option(
   "--follow-recommendation",
   "Automatically execute the recommended structural tool when Snipara returns one"
 ).option("--json", "Print raw JSON").action(async (options) => {
@@ -43953,6 +43980,7 @@ program.command("query").description(
     searchMode: options.searchMode,
     timeoutMs: parseInt(options.timeoutMs, 10),
     includeAnswerPack: options.answerPack !== false,
+    minimumChangeMode: resolveMinimumChangeMode(options.minimumChangeMode),
     autoDecompose: options.autoDecompose !== false,
     includeSharedContext: options.sharedContext !== false,
     followRecommendation: Boolean(options.followRecommendation),
@@ -45264,7 +45292,7 @@ collaboration.command("ide-status").description("Print compact JSON status for I
     json: options.json
   });
 });
-program.command("impact").description("Run bounded code impact with automatic local, hosted, or hybrid source selection").argument("[filePath]", "Source file to analyze").option("-q, --qualified-name <qualifiedName>", "Qualified symbol name").option("--symbol-key <symbolKey>", "Stable graph symbol key").option("-f, --file-path <filePath>", "Source file to analyze").option("--changed-files <changedFiles...>", "Changed files to analyze").option("--diff-summary <diffSummary>", "Natural-language summary of the change").option("-d, --depth <number>", "Transitive traversal depth", "3").option("--transitive", "Enable transitive traversal (enabled by default)").option("--direction <direction>", "in|out|both", "both").option("-e, --edge-kinds <edgeKinds...>", "Edge kinds to traverse").option("--max-nodes <number>", "Maximum graph nodes to visit", "500").option("-l, --limit <number>", "Maximum impact entries", "50").option("--source <source>", "auto|local|hosted|hybrid", "auto").option("--fallback-hosted", "Augment the local result with hosted graph context").option("--cached", "When local is selected, use the cached overlay if present").option("--max-files <number>", "Maximum supported code files for local overlay", "2000").option("--json", "Print raw JSON").action(async (filePath, options) => {
+program.command("impact").description("Run bounded code impact with automatic local, hosted, or hybrid source selection").argument("[filePath]", "Source file to analyze").option("-q, --qualified-name <qualifiedName>", "Qualified symbol name").option("--symbol-key <symbolKey>", "Stable graph symbol key").option("-f, --file-path <filePath>", "Source file to analyze").option("--changed-files <changedFiles...>", "Changed files to analyze").option("--diff-summary <diffSummary>", "Natural-language summary of the change").option("-d, --depth <number>", "Transitive traversal depth", "3").option("--transitive", "Enable transitive traversal (enabled by default)").option("--direction <direction>", "in|out|both", "both").option("-e, --edge-kinds <edgeKinds...>", "Edge kinds to traverse").option("--max-nodes <number>", "Maximum graph nodes to visit", "500").option("-l, --limit <number>", "Maximum impact entries", "50").option("--source <source>", "auto|local|hosted|hybrid", "auto").option("--minimum-change-mode <mode>", "off|review (opt-in non-blocking review)", "off").option("--fallback-hosted", "Augment the local result with hosted graph context").option("--cached", "When local is selected, use the cached overlay if present").option("--max-files <number>", "Maximum supported code files for local overlay", "2000").option("--json", "Print raw JSON").action(async (filePath, options) => {
   await codeGraphAutoSourceCommand("impact", {
     qualifiedName: options.qualifiedName,
     symbolKey: options.symbolKey,
@@ -45278,6 +45306,7 @@ program.command("impact").description("Run bounded code impact with automatic lo
     maxNodes: parseInt(options.maxNodes, 10),
     limit: parseInt(options.limit, 10),
     source: options.source,
+    minimumChangeMode: resolveMinimumChangeMode(options.minimumChangeMode),
     fallbackHosted: Boolean(options.fallbackHosted),
     cached: Boolean(options.cached),
     maxFiles: parseInt(options.maxFiles, 10),
@@ -45529,7 +45558,7 @@ code.addCommand(
     });
   })
 ).addCommand(
-  new import_commander.Command("impact").description("Run the primary agent-ready code impact gate with automatic source selection").argument("[filePath]", "Source file to analyze").option("-q, --qualified-name <qualifiedName>", "Qualified symbol name").option("--symbol-key <symbolKey>", "Stable graph symbol key").option("-f, --file-path <filePath>", "Source file to analyze").option("--changed-files <changedFiles...>", "Changed files to analyze").option("--diff-summary <diffSummary>", "Natural-language summary of the change").option("-d, --depth <number>", "Transitive traversal depth", "3").option("--transitive", "Enable transitive traversal (enabled by default)").option("--direction <direction>", "in|out|both", "both").option("-e, --edge-kinds <edgeKinds...>", "Edge kinds to traverse").option("--max-nodes <number>", "Maximum graph nodes to visit", "500").option("-l, --limit <number>", "Maximum impact entries", "50").option("--source <source>", "auto|local|hosted|hybrid", "auto").option("--fallback-hosted", "Augment the local result with hosted graph context").option("--cached", "When local is selected, use the cached overlay if present").option("--max-files <number>", "Maximum supported code files for local overlay", "2000").option("--json", "Print raw JSON").action(async (filePath, options) => {
+  new import_commander.Command("impact").description("Run the primary agent-ready code impact gate with automatic source selection").argument("[filePath]", "Source file to analyze").option("-q, --qualified-name <qualifiedName>", "Qualified symbol name").option("--symbol-key <symbolKey>", "Stable graph symbol key").option("-f, --file-path <filePath>", "Source file to analyze").option("--changed-files <changedFiles...>", "Changed files to analyze").option("--diff-summary <diffSummary>", "Natural-language summary of the change").option("-d, --depth <number>", "Transitive traversal depth", "3").option("--transitive", "Enable transitive traversal (enabled by default)").option("--direction <direction>", "in|out|both", "both").option("-e, --edge-kinds <edgeKinds...>", "Edge kinds to traverse").option("--max-nodes <number>", "Maximum graph nodes to visit", "500").option("-l, --limit <number>", "Maximum impact entries", "50").option("--source <source>", "auto|local|hosted|hybrid", "auto").option("--minimum-change-mode <mode>", "off|review (opt-in non-blocking review)", "off").option("--fallback-hosted", "Augment the local result with hosted graph context").option("--cached", "When local is selected, use the cached overlay if present").option("--max-files <number>", "Maximum supported code files for local overlay", "2000").option("--json", "Print raw JSON").action(async (filePath, options) => {
     await codeGraphAutoSourceCommand("impact", {
       qualifiedName: options.qualifiedName,
       symbolKey: options.symbolKey,
@@ -45543,6 +45572,7 @@ code.addCommand(
       maxNodes: parseInt(options.maxNodes, 10),
       limit: parseInt(options.limit, 10),
       source: options.source,
+      minimumChangeMode: resolveMinimumChangeMode(options.minimumChangeMode),
       fallbackHosted: Boolean(options.fallbackHosted),
       cached: Boolean(options.cached),
       maxFiles: parseInt(options.maxFiles, 10),
@@ -46093,6 +46123,7 @@ if (require.main === module) {
   resolveFeatureArtifactPaths,
   resolveFullWorkflowTokenBudget,
   resolveLocalWorkerRoutingDefaults,
+  resolveMinimumChangeMode,
   resolveQueryFromToolInput,
   retrieveContextPack,
   runMemoryGuardCheck,

package/docs/FULL_REFERENCE.md

@@ -1117,8 +1117,30 @@ durable memory policy locally.
 These commands keep local workflow state moving and call hosted Snipara only
 where the specific command needs hosted context or memory:
 
+#### Canonical command forms
+
+`workflow run --mode` accepts exactly these values:
+
+| CLI value     | Guide label         | Runtime behavior                                                      |
+| ------------- | ------------------- | --------------------------------------------------------------------- |
+| `lite`        | LITE                | Small, known-scope work with no mandatory hosted context call         |
+| `standard`    | STANDARD            | Normal work with context and code-graph follow-up when needed         |
+| `auto`        | AUTO                | Routes by task intent to `lite`, `standard`, `full`, or `orchestrate` |
+| `full`        | FULL                | Managed, phased work with durable context and plan support            |
+| `orchestrate` | FULL + ORCHESTRATED | Explicit deeper orchestration for multi-agent or proof-gate work      |
+
+The root `run` command is the Project Intelligence judgment/release flow.
+`workflow run` is the workflow-mode runner; they are different commands.
+Use root `final-commit` as the canonical final closeout. The registered
+`workflow final-commit` spelling is a compatibility alias. Similarly,
+`code impact` is the canonical impact gate, root `impact` is its compatibility
+alias, and `code local impact` is the separate repository-local overlay query.
+`task-commit` captures a durable task outcome; `workflow phase-commit`
+records one managed phase and advances the workflow.
+
 ```bash
 npx -y snipara-companion@latest workflow run --mode standard --query "who imports src.mcp_transport"
+npx -y snipara-companion@latest workflow run --mode auto --query "map the next bounded change"
 npx -y snipara-companion@latest workflow run --mode full --include-session-context --query "plan the auth refactor"
 npx -y snipara-companion@latest plan --query "plan the auth refactor" --write-plan-file .snipara/workflow/plans/auth-refactor-plan.json
 npx -y snipara-companion@latest task-commit --summary "Shipped auth refactor" --files apps/web/src/lib/auth.ts
@@ -1152,6 +1174,7 @@ snipara-companion workflow producer-review --artifact producer-abc123 --outcome
 snipara-companion final-commit --summary "Shipped auth hardening and tests" --why "Close the reported session replay gap" --evidence "passed:pnpm test auth" --next-step "Review the first 24 hours of telemetry" --files src/auth.ts tests/auth.test.ts
 snipara-companion doctor
 snipara-companion doctor --json
+snipara-companion collaboration guard --profile pre-deploy --enforce
 snipara-companion code callers --qualified-name src.mcp_transport.handle_call_tool
 snipara-companion code imports --file-path src/mcp_transport.py
 snipara-companion code neighbors --qualified-name src.mcp_transport.handle_call_tool --depth 3
@@ -1675,8 +1698,8 @@ Semantics:
 - `snipara-companion team-sync what-changed` = prints the local state summary and the hosted What Changed For Me response when configured
 - `snipara-companion team-sync sweep` = archives stale local work items after an inactivity threshold; default is 14 days and `--dry-run` previews candidates, actual archive count, and remaining stale work
 - `snipara-companion team-sync resume` = reloads local carryover plus the hosted latest handoff and checkpoint-aware resume guidance when available
-- `snipara-companion final-commit` / `workflow final-commit` = final hosted handoff plus a redacted seven-section closeout report in human output and `.snipara/workflow/final-report.json`; stored phase outcomes, pending Why Capture candidates, non-persisted items, evidence statuses, risks, and the next step remain visibly distinct
-- `snipara-companion code callers/imports/neighbors/shortest-path/impact` = primary code graph surface for agents with shell access. These commands use `--source auto` by default; clean configured checkouts use hosted MCP, dirty/ahead worktrees use a hosted-base plus local-delta hybrid, and unconfigured projects stay local. Every response reports `sourceSelection` and provenance.
+- `snipara-companion final-commit` (canonical) / `workflow final-commit` (compatibility alias) = final hosted handoff plus a redacted seven-section closeout report in human output and `.snipara/workflow/final-report.json`; stored phase outcomes, pending Why Capture candidates, non-persisted items, evidence statuses, risks, and the next step remain visibly distinct
+- `snipara-companion code callers/imports/neighbors/shortest-path/impact` = primary code graph surface for agents with shell access. These commands use `--source auto` by default; clean configured checkouts use hosted MCP, dirty/ahead worktrees use a hosted-base plus local-delta hybrid, and unconfigured projects stay local. Every response reports `sourceSelection` and provenance. The canonical impact spelling is `code impact`; root `impact` remains an alias.
 - `snipara-companion code symbol-card` = direct `snipara_code_symbol_card` for an important symbol before editing, with an agent guidance summary before raw JSON
 - `snipara-companion code impact --source hosted|local|hybrid` = optional source override. `--fallback-hosted` augments a local query when hosted auth is available; failures remain explicit degraded-local results.
 - `snipara-companion code local impact` = explicit repository-local bounded transitive impact. TypeScript uses compiler-AST calls/references/imports; Python and Go use import fallback. Use `--depth`, `--direction`, `--edge-kinds`, and `--max-nodes` to control expansion.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "snipara-companion",
-  "version": "3.6.13",
+  "version": "3.6.14",
   "description": "Local-first CLI that asks your repo what breaks before an AI coding agent edits it.",
   "main": "dist/index.js",
   "bin": {