node-autotask 0.90.0 → 0.92.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (42) hide show
  1. package/CHANGELOG.md +8 -0
  2. package/MCP_TOOL_CATALOG.json +360 -9
  3. package/README.md +1 -1
  4. package/capabilities.json +2022 -655
  5. package/dist/capabilities-data.js +2021 -654
  6. package/dist/capabilities.cjs +4 -4
  7. package/dist/capabilities.js +1 -1
  8. package/dist/{chunk-K5X4BPQC.js → chunk-2MC3LDZU.js} +3 -3
  9. package/dist/{chunk-4CI27SCM.cjs → chunk-ATDGRAX2.cjs} +33 -30
  10. package/dist/{chunk-HU6OGAPI.js → chunk-ELAU2MT5.js} +5 -2
  11. package/dist/{chunk-OI6E5XQA.cjs → chunk-H24SV774.cjs} +473 -122
  12. package/dist/{chunk-D7MJC7HA.cjs → chunk-JPZ7M7QI.cjs} +200 -200
  13. package/dist/{chunk-CY6USWUY.js → chunk-NQNAEA23.js} +468 -117
  14. package/dist/{chunk-ZLMYYO7R.cjs → chunk-QXJQIZCT.cjs} +3 -3
  15. package/dist/{chunk-Q73VQSUP.js → chunk-XAOWO42G.js} +2 -2
  16. package/dist/{index-4jmitXBs.d.cts → index-B9y5IwT1.d.cts} +115 -1
  17. package/dist/{index-4jmitXBs.d.ts → index-B9y5IwT1.d.ts} +115 -1
  18. package/dist/{index-DWyJAad_.d.ts → index-U7Lepdm2.d.ts} +10 -5
  19. package/dist/{index-BEqbDuMk.d.cts → index-lWek2NrV.d.cts} +10 -5
  20. package/dist/index.cjs +85 -85
  21. package/dist/index.d.cts +3 -3
  22. package/dist/index.d.ts +3 -3
  23. package/dist/index.js +5 -5
  24. package/dist/operations/index.cjs +51 -51
  25. package/dist/operations/index.d.cts +6 -4
  26. package/dist/operations/index.d.ts +6 -4
  27. package/dist/operations/index.js +2 -2
  28. package/dist/pool.cjs +12 -12
  29. package/dist/pool.d.cts +1 -1
  30. package/dist/pool.d.ts +1 -1
  31. package/dist/pool.js +4 -4
  32. package/dist/resources/index.cjs +27 -27
  33. package/dist/resources/index.d.cts +2 -2
  34. package/dist/resources/index.d.ts +2 -2
  35. package/dist/resources/index.js +3 -3
  36. package/dist/types/index.d.cts +1 -1
  37. package/dist/types/index.d.ts +1 -1
  38. package/docs/API.md +5 -2
  39. package/examples/mcp/manifest.ts +1 -1
  40. package/examples/mcp/tools-mode.ts +15 -6
  41. package/examples/mcp/tools-read.ts +8 -2
  42. package/package.json +1 -1
package/CHANGELOG.md CHANGED
@@ -7,6 +7,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
7
7
 
8
8
  ## [Unreleased]
9
9
 
10
+ ## [0.92.0] - 2026-11-21
11
+
12
+ - Add InventoryItems global reads and metadata, guarded create and collection PATCH, and fresh product/location and quantity checks. Replacing PUT and serial-number workflows remain tracked (#509, #468).
13
+
14
+ ## [0.91.0] - 2026-11-21
15
+
16
+ - Add read-only InternalLocations and ShippingTypes global query, count, item and metadata operations. Internal location country filters preserve the vendor’s literal `Other` rule for nonstandard countries; shipping billing-code navigation remains unspecified (#468).
17
+
10
18
  ## [0.90.0] - 2026-11-21
11
19
 
12
20
  ### Added
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "generatedBy": "scripts/build-tool-catalog.mjs",
3
- "generatedAt": "2026-11-21T03:40:53.270Z",
4
- "planHash": "2f78ad6e25fd96f024dfc749c3c9d1b8dbf52d8a23e097bd5046605a9975787c",
3
+ "generatedAt": "2026-11-21T07:49:36.813Z",
4
+ "planHash": "21e75d0f8bf9986b21c6b7996dfb1fec78b3aaa9afa86f50f5351c54c733d4ac",
5
5
  "metaTools": [
6
6
  {
7
7
  "name": "autotask_catalog",
@@ -51,7 +51,7 @@
51
51
  {
52
52
  "name": "autotask_read",
53
53
  "title": "Read One Registry Operation (read mode)",
54
- "description": "Route an effect=read registry operation by its exact name (never fuzzy, never a tool name): the 441 field/entity/UDF-info reads (getEntityInfo/getFieldInfo/getUdfInfo) have no dedicated tool and ride this one, and any read op is routable. For field discovery use operations.fieldCatalog (compact) instead of per-resource getFieldInfo (full dump). Read-only by construction — every op it can route is effect=read (zero mutating requests), and a write or delete op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). Reads have no dry run (a read with dry_run: true is refused). Harness approval policy: approve every tool with readOnlyHint !== true (exactly 291 today); the 674 readOnlyHint: true tools — including this one — are safe ungated; the 72 destructiveHint: true tools (71 dedicated delete tools + autotask_delete) require confirm === the operation key. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. For query/count ops args[0] is a params object (filter/limit); call autotask_describe first for the exact arg shape and filter grammar. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
54
+ "description": "Route an effect=read registry operation by its exact name (never fuzzy, never a tool name): the 450 field/entity/UDF-info reads (getEntityInfo/getFieldInfo/getUdfInfo) have no dedicated tool and ride this one, and any read op is routable. For field discovery use operations.fieldCatalog (compact) instead of per-resource getFieldInfo (full dump). Read-only by construction — every op it can route is effect=read (zero mutating requests), and a write or delete op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). Reads have no dry run (a read with dry_run: true is refused). Harness approval policy: approve every tool with readOnlyHint !== true (exactly 293 today); the 683 readOnlyHint: true tools — including this one — are safe ungated; the 72 destructiveHint: true tools (71 dedicated delete tools + autotask_delete) require confirm === the operation key. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. For query/count ops args[0] is a params object (filter/limit); call autotask_describe first for the exact arg shape and filter grammar. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
55
55
  "input": {
56
56
  "operation": {
57
57
  "type": "string",
@@ -71,7 +71,7 @@
71
71
  {
72
72
  "name": "autotask_write",
73
73
  "title": "Write One Registry Operation (write mode)",
74
- "description": "Route an effect=write registry operation by its exact name (never fuzzy, never a tool name), dry-run-first: the first call always travels dry_run: true (the pinned DryRunResult envelope, simulated: true — the mutating request is never issued); pass dry_run: false to apply. A read or delete op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). The 2 approval-gated writes (configuration_items.moveConfigurationItem, contacts.moveToCompany) additionally require confirm === the operation key, in BOTH dry-run and apply mode. Harness approval policy: this tool is one of the 291 tools that require approval (readOnlyHint !== true); the 674 readOnlyHint: true tools are safe ungated, and the 72 destructiveHint: true tools (71 dedicated delete tools + autotask_delete) require confirm === the operation key. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. For query/count ops args[0] is a params object (filter/limit); call autotask_describe first for the exact arg shape and filter grammar. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
74
+ "description": "Route an effect=write registry operation by its exact name (never fuzzy, never a tool name), dry-run-first: the first call always travels dry_run: true (the pinned DryRunResult envelope, simulated: true — the mutating request is never issued); pass dry_run: false to apply. A read or delete op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). The 2 approval-gated writes (configuration_items.moveConfigurationItem, contacts.moveToCompany) additionally require confirm === the operation key, in BOTH dry-run and apply mode. Harness approval policy: this tool is one of the 293 tools that require approval (readOnlyHint !== true); the 683 readOnlyHint: true tools are safe ungated, and the 72 destructiveHint: true tools (71 dedicated delete tools + autotask_delete) require confirm === the operation key. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. For query/count ops args[0] is a params object (filter/limit); call autotask_describe first for the exact arg shape and filter grammar. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
75
75
  "input": {
76
76
  "operation": {
77
77
  "type": "string",
@@ -101,7 +101,7 @@
101
101
  {
102
102
  "name": "autotask_delete",
103
103
  "title": "Delete One Registry Operation (delete mode)",
104
- "description": "Route an effect=destructive (delete) registry operation by its exact name (never fuzzy, never a tool name), dry-run-first: the first call always travels dry_run: true (the pinned DryRunResult envelope, simulated: true); pass dry_run: false to apply. confirm === the operation key is REQUIRED for every delete (all 71), in BOTH dry-run and apply mode — the deliberate act is the call, not the delete. A read or write op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). Harness approval policy: this tool is one of the 72 destructiveHint: true tools that require approval + confirm; approve every tool with readOnlyHint !== true (exactly 291 today); the 674 readOnlyHint: true tools are safe ungated. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
104
+ "description": "Route an effect=destructive (delete) registry operation by its exact name (never fuzzy, never a tool name), dry-run-first: the first call always travels dry_run: true (the pinned DryRunResult envelope, simulated: true); pass dry_run: false to apply. confirm === the operation key is REQUIRED for every delete (all 71), in BOTH dry-run and apply mode — the deliberate act is the call, not the delete. A read or write op is refused BEFORE any request (cross-mode refusal; the message names the correct tool). Harness approval policy: this tool is one of the 72 destructiveHint: true tools that require approval + confirm; approve every tool with readOnlyHint !== true (exactly 293 today); the 683 readOnlyHint: true tools are safe ungated. Migration: autotask_invoke is REPLACED by autotask_read/autotask_write/autotask_delete — update harness config. The argument set is CLOSED at the top level: an unknown argument is refused before any request, with a typed error naming the key and the valid keys.",
105
105
  "input": {
106
106
  "operation": {
107
107
  "type": "string",
@@ -4961,6 +4961,110 @@
4961
4961
  "reachReason": null,
4962
4962
  "boundedAlternative": null
4963
4963
  },
4964
+ {
4965
+ "operation": "internal_locations.count",
4966
+ "core": false,
4967
+ "tool": "autotask_internalLocation_count",
4968
+ "purpose": "Count internal location records matching a JSON filter.",
4969
+ "preferredWhen": null,
4970
+ "backingOperation": "internal_locations.count",
4971
+ "effect": "read",
4972
+ "dryRunFirst": false,
4973
+ "confirm": false,
4974
+ "requiredArgs": [],
4975
+ "tier": "extended",
4976
+ "reachable": true,
4977
+ "reach": "dedicated tool",
4978
+ "reachReason": null,
4979
+ "boundedAlternative": null
4980
+ },
4981
+ {
4982
+ "operation": "internal_locations.get",
4983
+ "core": false,
4984
+ "tool": "autotask_internalLocation_get",
4985
+ "purpose": "Get one InternalLocation by id.",
4986
+ "preferredWhen": "Use a known internal location id; use internal_locations.query for bounded discovery.",
4987
+ "backingOperation": "internal_locations.get",
4988
+ "effect": "read",
4989
+ "dryRunFirst": false,
4990
+ "confirm": false,
4991
+ "requiredArgs": [
4992
+ "id"
4993
+ ],
4994
+ "tier": "extended",
4995
+ "reachable": true,
4996
+ "reach": "dedicated tool",
4997
+ "reachReason": null,
4998
+ "boundedAlternative": null
4999
+ },
5000
+ {
5001
+ "operation": "internal_locations.getEntityInfo",
5002
+ "core": false,
5003
+ "tool": null,
5004
+ "purpose": "Get the InternalLocation capability metadata (canCreate/canUpdate/hasUserDefinedFields/...).",
5005
+ "preferredWhen": null,
5006
+ "backingOperation": "internal_locations.getEntityInfo",
5007
+ "effect": "read",
5008
+ "dryRunFirst": false,
5009
+ "confirm": false,
5010
+ "requiredArgs": [],
5011
+ "tier": "extended",
5012
+ "reachable": true,
5013
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
5014
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
5015
+ "boundedAlternative": null
5016
+ },
5017
+ {
5018
+ "operation": "internal_locations.getFieldInfo",
5019
+ "core": false,
5020
+ "tool": null,
5021
+ "purpose": "Get the InternalLocation field catalog (types, picklists, required, maxLength) — the full raw field dump (the largest read of the info trio; compact discovery: operations.fieldCatalog).",
5022
+ "preferredWhen": "Read-only internal location view configured in Admin Company Setup, distinct from CompanySiteConfigurations and InternalLocationWithBusinessHours; no create, update, delete or UDF writes, and no automatic write routing to another entity. timeZone returns a Microsoft Time Zone Index value. Country queries match the country display label; for a custom country mapped to Other, filter country with the literal Other rather than the custom text returned in the record. Visibility depends on API-user permissions and tenant data; an empty query does not prove absence. Enable redactPersonalData to mask the SDK personal-data field set.",
5023
+ "backingOperation": "internal_locations.getFieldInfo",
5024
+ "effect": "read",
5025
+ "dryRunFirst": false,
5026
+ "confirm": false,
5027
+ "requiredArgs": [],
5028
+ "tier": "extended",
5029
+ "reachable": true,
5030
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
5031
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
5032
+ "boundedAlternative": null
5033
+ },
5034
+ {
5035
+ "operation": "internal_locations.getUdfInfo",
5036
+ "core": false,
5037
+ "tool": null,
5038
+ "purpose": "Get the InternalLocation user-defined-field catalog.",
5039
+ "preferredWhen": null,
5040
+ "backingOperation": "internal_locations.getUdfInfo",
5041
+ "effect": "read",
5042
+ "dryRunFirst": false,
5043
+ "confirm": false,
5044
+ "requiredArgs": [],
5045
+ "tier": "extended",
5046
+ "reachable": true,
5047
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
5048
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
5049
+ "boundedAlternative": null
5050
+ },
5051
+ {
5052
+ "operation": "internal_locations.query",
5053
+ "core": false,
5054
+ "tool": "autotask_internalLocation_query",
5055
+ "purpose": "Query InternalLocation records with a JSON filter.",
5056
+ "preferredWhen": "Default for filtered reads (the bounded query op). queryAll is the unbounded raw walk — only when you need the ENTIRE set and accept a large payload; default to the bounded query op. Prefer an includeFields subset over the full model (field names via operations.fieldCatalog). latest: N is the latest-N rows (id-descending) when the newest rows are the goal (the vendor has no sort — count-anchored, docs/API.md §4). Prefer a hand-listed includeFields for common narrow reads (removed the closed fieldsPreset presets — the v1 billing_join field list for billing_items.query, if you need it, is: companyID, ticketID, projectID, timeEntryID, contractServiceID, contractID, billingCodeID, expenseItemID, itemDate, quantity, rate, extendedPrice, totalAmount, nonBillable); noteChars: N caps the untrusted-class free-text fields of the resource (notes/descriptions/UDF prose) to N chars each with the explicit … [+N chars truncated] marker (opt-in; the raw result when absent).",
5057
+ "backingOperation": "internal_locations.query",
5058
+ "effect": "read",
5059
+ "dryRunFirst": false,
5060
+ "confirm": false,
5061
+ "requiredArgs": [],
5062
+ "tier": "extended",
5063
+ "reachable": true,
5064
+ "reach": "dedicated tool",
5065
+ "reachReason": null,
5066
+ "boundedAlternative": null
5067
+ },
4964
5068
  {
4965
5069
  "operation": "inventory_products.count",
4966
5070
  "core": false,
@@ -10182,6 +10286,110 @@
10182
10286
  "reachReason": null,
10183
10287
  "boundedAlternative": null
10184
10288
  },
10289
+ {
10290
+ "operation": "shipping_types.count",
10291
+ "core": false,
10292
+ "tool": "autotask_shippingType_count",
10293
+ "purpose": "Count shipping type records matching a JSON filter.",
10294
+ "preferredWhen": null,
10295
+ "backingOperation": "shipping_types.count",
10296
+ "effect": "read",
10297
+ "dryRunFirst": false,
10298
+ "confirm": false,
10299
+ "requiredArgs": [],
10300
+ "tier": "extended",
10301
+ "reachable": true,
10302
+ "reach": "dedicated tool",
10303
+ "reachReason": null,
10304
+ "boundedAlternative": null
10305
+ },
10306
+ {
10307
+ "operation": "shipping_types.get",
10308
+ "core": false,
10309
+ "tool": "autotask_shippingType_get",
10310
+ "purpose": "Get one ShippingType by id.",
10311
+ "preferredWhen": "Use a known shipping type id; use shipping_types.query for bounded discovery.",
10312
+ "backingOperation": "shipping_types.get",
10313
+ "effect": "read",
10314
+ "dryRunFirst": false,
10315
+ "confirm": false,
10316
+ "requiredArgs": [
10317
+ "id"
10318
+ ],
10319
+ "tier": "extended",
10320
+ "reachable": true,
10321
+ "reach": "dedicated tool",
10322
+ "reachReason": null,
10323
+ "boundedAlternative": null
10324
+ },
10325
+ {
10326
+ "operation": "shipping_types.getEntityInfo",
10327
+ "core": false,
10328
+ "tool": null,
10329
+ "purpose": "Get the ShippingType capability metadata (canCreate/canUpdate/hasUserDefinedFields/...).",
10330
+ "preferredWhen": null,
10331
+ "backingOperation": "shipping_types.getEntityInfo",
10332
+ "effect": "read",
10333
+ "dryRunFirst": false,
10334
+ "confirm": false,
10335
+ "requiredArgs": [],
10336
+ "tier": "extended",
10337
+ "reachable": true,
10338
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
10339
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
10340
+ "boundedAlternative": null
10341
+ },
10342
+ {
10343
+ "operation": "shipping_types.getFieldInfo",
10344
+ "core": false,
10345
+ "tool": null,
10346
+ "purpose": "Get the ShippingType field catalog (types, picklists, required, maxLength) — the full raw field dump (the largest read of the info trio; compact discovery: operations.fieldCatalog).",
10347
+ "preferredWhen": "Read-only shipping catalog configured in Admin and referenced by Quotes in CRM; no create, update, delete or UDF writes. Visibility depends on API-user permissions and tenant data; an empty query does not prove absence. The official field table labels billingCodeID as a ProjectCharges reference, but its name and that reference disagree; no new navigation contract is inferred. Discover authoritative field metadata with shipping_types.getFieldInfo before interpreting the reference.",
10348
+ "backingOperation": "shipping_types.getFieldInfo",
10349
+ "effect": "read",
10350
+ "dryRunFirst": false,
10351
+ "confirm": false,
10352
+ "requiredArgs": [],
10353
+ "tier": "extended",
10354
+ "reachable": true,
10355
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
10356
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
10357
+ "boundedAlternative": null
10358
+ },
10359
+ {
10360
+ "operation": "shipping_types.getUdfInfo",
10361
+ "core": false,
10362
+ "tool": null,
10363
+ "purpose": "Get the ShippingType user-defined-field catalog.",
10364
+ "preferredWhen": null,
10365
+ "backingOperation": "shipping_types.getUdfInfo",
10366
+ "effect": "read",
10367
+ "dryRunFirst": false,
10368
+ "confirm": false,
10369
+ "requiredArgs": [],
10370
+ "tier": "extended",
10371
+ "reachable": true,
10372
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
10373
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
10374
+ "boundedAlternative": null
10375
+ },
10376
+ {
10377
+ "operation": "shipping_types.query",
10378
+ "core": false,
10379
+ "tool": "autotask_shippingType_query",
10380
+ "purpose": "Query ShippingType records with a JSON filter.",
10381
+ "preferredWhen": "Default for filtered reads (the bounded query op). queryAll is the unbounded raw walk — only when you need the ENTIRE set and accept a large payload; default to the bounded query op. Prefer an includeFields subset over the full model (field names via operations.fieldCatalog). latest: N is the latest-N rows (id-descending) when the newest rows are the goal (the vendor has no sort — count-anchored, docs/API.md §4). Prefer a hand-listed includeFields for common narrow reads (removed the closed fieldsPreset presets — the v1 billing_join field list for billing_items.query, if you need it, is: companyID, ticketID, projectID, timeEntryID, contractServiceID, contractID, billingCodeID, expenseItemID, itemDate, quantity, rate, extendedPrice, totalAmount, nonBillable); noteChars: N caps the untrusted-class free-text fields of the resource (notes/descriptions/UDF prose) to N chars each with the explicit … [+N chars truncated] marker (opt-in; the raw result when absent).",
10382
+ "backingOperation": "shipping_types.query",
10383
+ "effect": "read",
10384
+ "dryRunFirst": false,
10385
+ "confirm": false,
10386
+ "requiredArgs": [],
10387
+ "tier": "extended",
10388
+ "reachable": true,
10389
+ "reach": "dedicated tool",
10390
+ "reachReason": null,
10391
+ "boundedAlternative": null
10392
+ },
10185
10393
  {
10186
10394
  "operation": "skills.count",
10187
10395
  "core": false,
@@ -19958,6 +20166,149 @@
19958
20166
  "dateConventions": "date convention (verify against current field metadata): startDate and endDate are date-only at 00:00Z; tickets.dueDateTime is a true UTC instant."
19959
20167
  }
19960
20168
  },
20169
+ {
20170
+ "operation": "inventory_items.count",
20171
+ "core": false,
20172
+ "tool": "autotask_inventoryItem_count",
20173
+ "purpose": "Count inventory item records matching a JSON filter.",
20174
+ "preferredWhen": null,
20175
+ "backingOperation": "inventory_items.count",
20176
+ "effect": "read",
20177
+ "dryRunFirst": false,
20178
+ "confirm": false,
20179
+ "requiredArgs": [],
20180
+ "tier": "extended",
20181
+ "reachable": true,
20182
+ "reach": "dedicated tool",
20183
+ "reachReason": null,
20184
+ "boundedAlternative": null
20185
+ },
20186
+ {
20187
+ "operation": "inventory_items.create",
20188
+ "core": false,
20189
+ "tool": "autotask_inventoryItem_create",
20190
+ "purpose": "Create a product-location inventory item.",
20191
+ "preferredWhen": "Use for a new product-location pair after checking known inventory items.",
20192
+ "backingOperation": "inventory_items.create",
20193
+ "effect": "write",
20194
+ "dryRunFirst": true,
20195
+ "confirm": false,
20196
+ "requiredArgs": [
20197
+ "data"
20198
+ ],
20199
+ "tier": "extended",
20200
+ "reachable": true,
20201
+ "reach": "dedicated tool",
20202
+ "reachReason": null,
20203
+ "boundedAlternative": null
20204
+ },
20205
+ {
20206
+ "operation": "inventory_items.get",
20207
+ "core": false,
20208
+ "tool": "autotask_inventoryItem_get",
20209
+ "purpose": "Get one InventoryItem by id.",
20210
+ "preferredWhen": "Preferred when the numeric id is already known.",
20211
+ "backingOperation": "inventory_items.get",
20212
+ "effect": "read",
20213
+ "dryRunFirst": false,
20214
+ "confirm": false,
20215
+ "requiredArgs": [
20216
+ "id"
20217
+ ],
20218
+ "tier": "extended",
20219
+ "reachable": true,
20220
+ "reach": "dedicated tool",
20221
+ "reachReason": null,
20222
+ "boundedAlternative": null
20223
+ },
20224
+ {
20225
+ "operation": "inventory_items.getEntityInfo",
20226
+ "core": false,
20227
+ "tool": null,
20228
+ "purpose": "Get the InventoryItem capability metadata (canCreate/canUpdate/hasUserDefinedFields/...).",
20229
+ "preferredWhen": null,
20230
+ "backingOperation": "inventory_items.getEntityInfo",
20231
+ "effect": "read",
20232
+ "dryRunFirst": false,
20233
+ "confirm": false,
20234
+ "requiredArgs": [],
20235
+ "tier": "extended",
20236
+ "reachable": true,
20237
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
20238
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
20239
+ "boundedAlternative": null
20240
+ },
20241
+ {
20242
+ "operation": "inventory_items.getFieldInfo",
20243
+ "core": false,
20244
+ "tool": null,
20245
+ "purpose": "Get the InventoryItem field catalog (types, picklists, required, maxLength) — the full raw field dump (the largest read of the info trio; compact discovery: operations.fieldCatalog).",
20246
+ "preferredWhen": null,
20247
+ "backingOperation": "inventory_items.getFieldInfo",
20248
+ "effect": "read",
20249
+ "dryRunFirst": false,
20250
+ "confirm": false,
20251
+ "requiredArgs": [],
20252
+ "tier": "extended",
20253
+ "reachable": true,
20254
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
20255
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
20256
+ "boundedAlternative": null
20257
+ },
20258
+ {
20259
+ "operation": "inventory_items.getUdfInfo",
20260
+ "core": false,
20261
+ "tool": null,
20262
+ "purpose": "Get the InventoryItem user-defined-field catalog.",
20263
+ "preferredWhen": null,
20264
+ "backingOperation": "inventory_items.getUdfInfo",
20265
+ "effect": "read",
20266
+ "dryRunFirst": false,
20267
+ "confirm": false,
20268
+ "requiredArgs": [],
20269
+ "tier": "extended",
20270
+ "reachable": true,
20271
+ "reach": "autotask_read (exact operation name; effect=read — read-only; zero mutating requests)",
20272
+ "reachReason": "no dedicated tool — the op is still discoverable and invocable by exact name",
20273
+ "boundedAlternative": null
20274
+ },
20275
+ {
20276
+ "operation": "inventory_items.query",
20277
+ "core": false,
20278
+ "tool": "autotask_inventoryItem_query",
20279
+ "purpose": "Query InventoryItem records with a JSON filter.",
20280
+ "preferredWhen": "Default for filtered reads (the bounded query op). queryAll is the unbounded raw walk — only when you need the ENTIRE set and accept a large payload; default to the bounded query op. Prefer an includeFields subset over the full model (field names via operations.fieldCatalog). latest: N is the latest-N rows (id-descending) when the newest rows are the goal (the vendor has no sort — count-anchored, docs/API.md §4). Prefer a hand-listed includeFields for common narrow reads (removed the closed fieldsPreset presets — the v1 billing_join field list for billing_items.query, if you need it, is: companyID, ticketID, projectID, timeEntryID, contractServiceID, contractID, billingCodeID, expenseItemID, itemDate, quantity, rate, extendedPrice, totalAmount, nonBillable); noteChars: N caps the untrusted-class free-text fields of the resource (notes/descriptions/UDF prose) to N chars each with the explicit … [+N chars truncated] marker (opt-in; the raw result when absent).",
20281
+ "backingOperation": "inventory_items.query",
20282
+ "effect": "read",
20283
+ "dryRunFirst": false,
20284
+ "confirm": false,
20285
+ "requiredArgs": [],
20286
+ "tier": "extended",
20287
+ "reachable": true,
20288
+ "reach": "dedicated tool",
20289
+ "reachReason": null,
20290
+ "boundedAlternative": null
20291
+ },
20292
+ {
20293
+ "operation": "inventory_items.update",
20294
+ "core": false,
20295
+ "tool": "autotask_inventoryItem_update",
20296
+ "purpose": "Update writable inventory item fields with collection PATCH.",
20297
+ "preferredWhen": "Use with a known item ID and only the writable fields that should change.",
20298
+ "backingOperation": "inventory_items.update",
20299
+ "effect": "write",
20300
+ "dryRunFirst": true,
20301
+ "confirm": false,
20302
+ "requiredArgs": [
20303
+ "id",
20304
+ "data"
20305
+ ],
20306
+ "tier": "extended",
20307
+ "reachable": true,
20308
+ "reach": "dedicated tool",
20309
+ "reachReason": null,
20310
+ "boundedAlternative": null
20311
+ },
19961
20312
  {
19962
20313
  "operation": "inventory_locations.count",
19963
20314
  "core": false,
@@ -25733,11 +26084,11 @@
25733
26084
  }
25734
26085
  ],
25735
26086
  "counts": {
25736
- "registryOperations": 1401,
26087
+ "registryOperations": 1421,
25737
26088
  "coreOperations": 10,
25738
- "catalogRows": 1401,
25739
- "withTool": 960,
25740
- "readModeOnly": 441,
26089
+ "catalogRows": 1421,
26090
+ "withTool": 971,
26091
+ "readModeOnly": 450,
25741
26092
  "neverExposed": 0,
25742
26093
  "metaTools": 5
25743
26094
  }
package/README.md CHANGED
@@ -388,7 +388,7 @@ The SDK ships the mechanical substrate, not a server (no MCP imports, no zod):
388
388
  governor place: `src/operations/invoke.ts`).
389
389
 
390
390
  <!-- BEGIN GENERATED:approval-policy -->
391
- **Harness approval policy (recommended).** Require approval on every tool whose MCP annotation is `readOnlyHint !== true` — today exactly **291 tools**: the 289 dedicated write/delete tools (218 `effect=write` + 71 `effect=destructive`, every one of them has a dedicated tool) + `autotask_write` + `autotask_delete`. The remaining **674 tools** (the 661 dedicated read tools + the 10 core read tools + `autotask_read` + `autotask_catalog` + `autotask_describe`) all carry `readOnlyHint: true` and are safe ungated: every op they can route is `effect=read` (zero mutating requests). Refined class for the highest-risk ops: the **72 tools with `destructiveHint: true`** (71 dedicated delete tools + `autotask_delete`) additionally require `confirm === <operation key>` even inside `autotask_delete`. The approval-gated writes (`configuration_items.moveConfigurationItem`, `contacts.moveToCompany`) require the same confirm inside `autotask_write`. Migration: `autotask_invoke` is REPLACED by `autotask_read`/`autotask_write`/`autotask_delete` — update harness config and remove any `autotask_invoke` entry. The rule is derivable from the registry `effect` column and the mode tools’ fixed annotations; it never needs a hand-maintained name list.
391
+ **Harness approval policy (recommended).** Require approval on every tool whose MCP annotation is `readOnlyHint !== true` — today exactly **293 tools**: the 291 dedicated write/delete tools (220 `effect=write` + 71 `effect=destructive`, every one of them has a dedicated tool) + `autotask_write` + `autotask_delete`. The remaining **683 tools** (the 670 dedicated read tools + the 10 core read tools + `autotask_read` + `autotask_catalog` + `autotask_describe`) all carry `readOnlyHint: true` and are safe ungated: every op they can route is `effect=read` (zero mutating requests). Refined class for the highest-risk ops: the **72 tools with `destructiveHint: true`** (71 dedicated delete tools + `autotask_delete`) additionally require `confirm === <operation key>` even inside `autotask_delete`. The approval-gated writes (`configuration_items.moveConfigurationItem`, `contacts.moveToCompany`) require the same confirm inside `autotask_write`. Migration: `autotask_invoke` is REPLACED by `autotask_read`/`autotask_write`/`autotask_delete` — update harness config and remove any `autotask_invoke` entry. The rule is derivable from the registry `effect` column and the mode tools’ fixed annotations; it never needs a hand-maintained name list.
392
392
  <!-- END GENERATED:approval-policy -->
393
393
  **Provider tool-search interop.** Hosts with provider-native tool search
394
394
  (OpenAI `type: "tool_search"` / `defer_loading`; Anthropic's tool search