mammoth-cli 2.2.57__tar.gz → 2.2.59__tar.gz

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mammoth-cli
-Version: 2.2.57
+Version: 2.2.59
 Summary: Command-line interface for the Mammoth Analytics platform
 Project-URL: homepage, https://mammoth.io
 Project-URL: repository, https://github.com/EdgeMetric/mammothsdk
@@ -17,7 +17,7 @@ Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Requires-Python: <3.15,>=3.12
 Requires-Dist: jsonschema>=4.20
 Requires-Dist: keyring<26,>=25.7
-Requires-Dist: mammoth-io<0.9,>=0.8.30
+Requires-Dist: mammoth-io<0.9,>=0.8.31
 Requires-Dist: platformdirs<5,>=4.10
 Requires-Dist: pydantic<3,>=2.12.5
 Requires-Dist: pyyaml<7,>=6

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/__init__.py

@@ -6,7 +6,7 @@ not implement a second HTTP client and does not call private SDK members.
 
 from __future__ import annotations
 
-__version__ = "2.2.57"
+__version__ = "2.2.59"
 
 # The versioned machine output envelope contract.
 SCHEMA_VERSION = 1

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/app.py

@@ -762,6 +762,18 @@ def _shared_option_params() -> list[inspect.Parameter]:
                 ),
             ],
         ),
+        opt(
+            "standing",
+            False,
+            Annotated[
+                bool,
+                typer.Option(
+                    "--standing",
+                    help="Stage a step that changes no row today, as a rule for future data.",
+                    rich_help_panel="Safety",
+                ),
+            ],
+        ),
         opt(
             "input_file",
             None,

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/bundled_skill/mammoth-cli/references/capabilities.md

@@ -1,7 +1,7 @@
 # What is proven on release
 
 Generated from `docs/release-capability-matrix.json`; do not edit by hand.
-The CLI publishes 656 commands. 516 of them bind one of the 545 API operations in the matrix; the remainder are local commands (`schema`, `auth`, `doctor`, `log`, ...) or typed variants that share an operation (`view transform *` commands submit through `view.task.add`, except `rename-columns` and `sort`, which set view display properties through `view.update`). 262 bound commands ran once successfully on release, 1 are not supported there, and the rest are untried. Untried is not broken: discover the contract with `mammoth schema get COMMAND_ID`, run it, and treat the structured error envelope as the answer. Use an untried command normally: confirm a write with the user first, check its result with a read afterwards, and never refuse a command only because it is untried.
+The CLI publishes 691 commands. 516 of them bind one of the 545 API operations in the matrix; the remainder are local commands (`schema`, `auth`, `doctor`, `log`, ...) or typed variants that share an operation (`view transform *` commands submit through `view.task.add`, except `rename-columns` and `sort`, which set view display properties through `view.update`). 262 bound commands ran once successfully on release, 1 are not supported there, and the rest are untried. Untried is not broken: discover the contract with `mammoth schema get COMMAND_ID`, run it, and treat the structured error envelope as the answer. Use an untried command normally: confirm a write with the user first, check its result with a read afterwards, and never refuse a command only because it is untried.
 
 Status meanings:
 

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/bundled_skill/mammoth-cli/references/command-index.md

@@ -1,12 +1,12 @@
 # Published CLI command catalog
 
-This generated catalog is a lookup table for every command in the published CLI manifest, currently 663; each entry carries its release status line. It is a command-contract index, not an assertion that every backend/OpenAPI operation has a CLI binding. For an exact local CLI contract, run `mammoth schema get COMMAND_ID`; `mammoth capability list` is an API-binding inventory and can omit typed/local CLI routes. For focused workflows, read the [recipes index](recipes/index.md). Sensitive structured input must come from a private file or pipe; never put secrets in literal argv.
+This generated catalog is a lookup table for every command in the published CLI manifest, currently 691; each entry carries its release status line. It is a command-contract index, not an assertion that every backend/OpenAPI operation has a CLI binding. For an exact local CLI contract, run `mammoth schema get COMMAND_ID`; `mammoth capability list` is an API-binding inventory and can omit typed/local CLI routes. For focused workflows, read the [recipes index](recipes/index.md). Sensitive structured input must come from a private file or pipe; never put secrets in literal argv.
 
 Load only the applicable domain file:
 
 - [activity](commands/activity.md) — 2 published commands
 - [addon](commands/addon.md) — 7 published commands
-- [agent](commands/agent.md) — 15 published commands
+- [agent](commands/agent.md) — 43 published commands
 - [ai](commands/ai.md) — 5 published commands
 - [annotation](commands/annotation.md) — 5 published commands
 - [auth](commands/auth.md) — 3 published commands

mammoth_cli-2.2.59/mammoth_cli/bundled_skill/mammoth-cli/references/commands/agent.md

@@ -0,0 +1,433 @@
+# `agent` commands
+
+Every command returns the standard JSON envelope. On nonzero exit, read the error envelope and its `recovery_commands`; do not guess request fields. "Status on release" is joined from `docs/release-capability-matrix.json`: *ran once* means one bounded live run succeeded on the named CLI release, *untried* means nobody has run it (use it normally: confirm writes with the user, check the result afterwards), *not supported* means the backend refuses it. When this file disagrees with [capabilities](../capabilities.md) or a recipe, they win. Envelope shapes: [machine output](../machine-output.md).
+
+### `agent.access.set`
+
+Run: `mammoth agent access set`. Exact input fields: `mammoth schema get agent.access.set`.
+
+Example: `mammoth agent access set sample --input '{"role": "member", "propose": true}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentAccessSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.action.delete`
+
+Run: `mammoth agent action delete`. Exact input fields: `mammoth schema get agent.action.delete`.
+
+Example: `mammoth agent action delete resource-123`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentActionDeleteResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.action.list`
+
+Run: `mammoth agent action list`. Exact input fields: `mammoth schema get agent.action.list`.
+
+Example: `mammoth agent action list`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentActionListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.charter.get`
+
+Run: `mammoth agent charter get`. Exact input fields: `mammoth schema get agent.charter.get`.
+
+Example: `mammoth agent charter get sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentCharterGetResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.charter.restore`
+
+Run: `mammoth agent charter restore`. Exact input fields: `mammoth schema get agent.charter.restore`.
+
+Example: `mammoth agent charter restore sample 123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentCharterRestoreResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.charter.set`
+
+Run: `mammoth agent charter set`. Exact input fields: `mammoth schema get agent.charter.set`.
+
+Example: `mammoth agent charter set sample --input '{"charter": "Watch weekly margin and report the three biggest drops."}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentCharterSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.charter.versions`
+
+Run: `mammoth agent charter versions`. Exact input fields: `mammoth schema get agent.charter.versions`.
+
+Example: `mammoth agent charter versions sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentCharterVersionsResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.chat`
+
+Run: `mammoth agent chat`. Exact input fields: `mammoth schema get agent.chat`.
+
+Example: `mammoth agent chat --input '{"message": "Summarize revenue by region", "scope": {"sample_key": "Status"}}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentChatResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.create`
+
+Run: `mammoth agent create`. Exact input fields: `mammoth schema get agent.create`.
+
+Example: `mammoth agent create Status --input '{"name": "Revenue report"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentCreateResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.delete`
+
+Run: `mammoth agent delete`. Exact input fields: `mammoth schema get agent.delete`.
+
+Example: `mammoth agent delete sample`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentDeleteResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.disable`
+
+Run: `mammoth agent disable`. Exact input fields: `mammoth schema get agent.disable`.
+
+Example: `mammoth agent disable sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentDisableResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.feedback.list`
+
+Run: `mammoth agent feedback list`. Exact input fields: `mammoth schema get agent.feedback.list`.
+
+Example: `mammoth agent feedback list sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentFeedbackListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.get`
+
+Run: `mammoth agent get`. Exact input fields: `mammoth schema get agent.get`.
+
+Example: `mammoth agent get sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentGetResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.goldens.add`
+
+Run: `mammoth agent goldens add`. Exact input fields: `mammoth schema get agent.goldens.add`.
+
+Example: `mammoth agent goldens add sample --input '{"question": "Summarize revenue by region", "expected": "sample"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentGoldensAddResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.goldens.list`
+
+Run: `mammoth agent goldens list`. Exact input fields: `mammoth schema get agent.goldens.list`.
+
+Example: `mammoth agent goldens list sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentGoldensListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.goldens.remove`
+
+Run: `mammoth agent goldens remove`. Exact input fields: `mammoth schema get agent.goldens.remove`.
+
+Example: `mammoth agent goldens remove sample 123`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentGoldensRemoveResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.goldens.run`
+
+Run: `mammoth agent goldens run`. Exact input fields: `mammoth schema get agent.goldens.run`.
+
+Example: `mammoth agent goldens run sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentGoldensRunResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.goldens.status`
+
+Run: `mammoth agent goldens status`. Exact input fields: `mammoth schema get agent.goldens.status`.
+
+Example: `mammoth agent goldens status sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentGoldensStatusResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.list`
+
+Run: `mammoth agent list`. Exact input fields: `mammoth schema get agent.list`.
+
+Example: `mammoth agent list`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.memory.add`
+
+Run: `mammoth agent memory add`. Exact input fields: `mammoth schema get agent.memory.add`.
+
+Example: `mammoth agent memory add sample --input '{"name": "Revenue report", "content": "sample"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentMemoryAddResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.memory.list`
+
+Run: `mammoth agent memory list`. Exact input fields: `mammoth schema get agent.memory.list`.
+
+Example: `mammoth agent memory list sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentMemoryListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.memory.remove`
+
+Run: `mammoth agent memory remove`. Exact input fields: `mammoth schema get agent.memory.remove`.
+
+Example: `mammoth agent memory remove sample 123`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentMemoryRemoveResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.projects.set`
+
+Run: `mammoth agent projects set`. Exact input fields: `mammoth schema get agent.projects.set`.
+
+Example: `mammoth agent projects set sample --input '{"project_ids": [12, 15]}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentProjectsSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.publish`
+
+Run: `mammoth agent publish`. Exact input fields: `mammoth schema get agent.publish`.
+
+Example: `mammoth agent publish sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentPublishResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.roles`
+
+Run: `mammoth agent roles`. Exact input fields: `mammoth schema get agent.roles`.
+
+Example: `mammoth agent roles`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRolesResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.extend`
+
+Run: `mammoth agent run extend`. Exact input fields: `mammoth schema get agent.run.extend`.
+
+Example: `mammoth agent run extend resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunExtendResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.list`
+
+Run: `mammoth agent run list`. Exact input fields: `mammoth schema get agent.run.list`.
+
+Example: `mammoth agent run list`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.pause`
+
+Run: `mammoth agent run pause`. Exact input fields: `mammoth schema get agent.run.pause`.
+
+Example: `mammoth agent run pause resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunPauseResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.resume`
+
+Run: `mammoth agent run resume`. Exact input fields: `mammoth schema get agent.run.resume`.
+
+Example: `mammoth agent run resume resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunResumeResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.status`
+
+Run: `mammoth agent run status`. Exact input fields: `mammoth schema get agent.run.status`.
+
+Example: `mammoth agent run status`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunStatusResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.stop`
+
+Run: `mammoth agent run stop`. Exact input fields: `mammoth schema get agent.run.stop`.
+
+Example: `mammoth agent run stop resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunStopResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.run.units.set`
+
+Run: `mammoth agent run units set`. Exact input fields: `mammoth schema get agent.run.units.set`.
+
+Example: `mammoth agent run units set resource-123 --input '{"step": 1, "kind": "sample", "units": [{"sample_key": "Status"}]}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentRunUnitsSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.scratch.clear`
+
+Run: `mammoth agent scratch clear`. Exact input fields: `mammoth schema get agent.scratch.clear`.
+
+Example: `mammoth agent scratch clear sample 'Revenue report'`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentScratchClearResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.scratch.get`
+
+Run: `mammoth agent scratch get`. Exact input fields: `mammoth schema get agent.scratch.get`.
+
+Example: `mammoth agent scratch get sample 'Revenue report'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentScratchGetResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.scratch.list`
+
+Run: `mammoth agent scratch list`. Exact input fields: `mammoth schema get agent.scratch.list`.
+
+Example: `mammoth agent scratch list sample`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentScratchListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.scratch.set`
+
+Run: `mammoth agent scratch set`. Exact input fields: `mammoth schema get agent.scratch.set`.
+
+Example: `mammoth agent scratch set sample 'Revenue report' --input '{"content": "sample"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentScratchSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.session.delete`
+
+Run: `mammoth agent session delete`. Exact input fields: `mammoth schema get agent.session.delete`.
+
+Example: `mammoth agent session delete resource-123`. Illustrative only: append `--yes` after observing an owned target.
+
+Result: `AgentSessionDeleteResult`; mutation `destructive`, confirmation `prompt_or_yes`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.session.list`
+
+Run: `mammoth agent session list`. Exact input fields: `mammoth schema get agent.session.list`.
+
+Example: `mammoth agent session list`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentSessionListResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: ran once on CLI 2.0.11 — Read-only sweep 2026-09-18: exit 0 on release with CLI 2.0.11; result keys: sessions. Single read only; no fixture variants, error envelopes, or write paths assessed.
+
+### `agent.session.messages`
+
+Run: `mammoth agent session messages`. Exact input fields: `mammoth schema get agent.session.messages`.
+
+Example: `mammoth agent session messages resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentSessionMessagesResult`; mutation `read`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.session.set-visibility`
+
+Run: `mammoth agent session set-visibility`. Exact input fields: `mammoth schema get agent.session.set-visibility`.
+
+Example: `mammoth agent session set-visibility resource-123 --input '{"visibility": "sample"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentSessionSetVisibilityResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.team.set`
+
+Run: `mammoth agent team set`. Exact input fields: `mammoth schema get agent.team.set`.
+
+Example: `mammoth agent team set sample --input '{"team": {"agents": ["margin-watch"], "max_rounds": 4}}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentTeamSetResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.turn.cancel`
+
+Run: `mammoth agent turn cancel`. Exact input fields: `mammoth schema get agent.turn.cancel`.
+
+Example: `mammoth agent turn cancel resource-123`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentTurnCancelResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.
+
+### `agent.update`
+
+Run: `mammoth agent update`. Exact input fields: `mammoth schema get agent.update`.
+
+Example: `mammoth agent update sample --input '{"name": "Margin watch"}'`. Placeholders are illustrative; resolve IDs and input from observed reads.
+
+Result: `AgentUpdateResult`; mutation `benign_mutation`, confirmation `none`, wait policy `not_async`.
+
+Status on release: untried; no live run recorded.

mammoth_cli-2.2.59/mammoth_cli/bundled_skill/mammoth-cli/references/recipes/build-an-agent.md

@@ -0,0 +1,31 @@
+# Building an agent
+
+The charter is the method: the job, the ordered steps, what counts as a
+finding, the report format and what the agent never does. Write it first.
+
+```bash
+mammoth agent roles
+mammoth agent create KEY --input '{"name": "Margin watch", "charter": "...", "role": "ROLE", "project_ids": [12]}'
+mammoth agent goldens add KEY --input '{"question": "...", "expected": "..."}'
+mammoth agent goldens run KEY
+mammoth agent goldens status KEY
+mammoth agent publish KEY
+```
+
+- Pick a built-in role from `agent roles`. A new agent proposes changes for
+  approval; keep that unless the owner asks otherwise
+  (`agent access set KEY --input '{"role": "ROLE", "propose": false}'`).
+- Add a golden for each question a right agent must answer, then run the
+  proof and read `agent goldens status` until it is finished. Failed goldens
+  are shown, not blocking: fix the charter (`agent charter set KEY --input
+  '{"charter": "..."}'`; `agent charter versions` and `agent charter restore`
+  go back) and run again. `agent publish` is refused without a finished run
+  for the current charter.
+- `agent projects set KEY --input '{"project_ids": [12, 15]}'` chooses where
+  it works. `agent team set KEY --input '{"team": {"agents": ["other-key"],
+  "max_rounds": 4}}'` lets it ask other published agents.
+- `agent feedback list KEY` shows the thumbs up and down people gave it.
+
+An agent running inside the app saves learned facts with `agent memory add
+KEY --project P --input '{"name": "...", "content": "..."}'` and working notes
+with `agent scratch set KEY NAME --project P --input '{"content": "..."}'`.

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/bundled_skill/mammoth-cli/references/recipes/index.md

@@ -18,6 +18,7 @@ hashes; stop on ambiguous schemas, authorization errors, or unknown effects.
 - [typed ETL discovery](typed-etl-discovery.md)
 - [exports and artifacts](exports.md)
 - [dashboards](dashboards.md)
+- [build, prove and publish an agent](build-an-agent.md)
 - [recurring work: automations and schedules](scheduling.md)
 - [trash, recovery and cleanup](cleanup.md)
 - [deliverable retention and cleanup authorization](../retention.md)

{mammoth_cli-2.2.57 → mammoth_cli-2.2.59}/mammoth_cli/bundled_skill/mammoth-cli/references/recipes/transforms.md

@@ -90,7 +90,8 @@ states `duplicates` (`exact_duplicate_rows` over `rows_checked` of `row_count`;
 on a later page it names the table-wide check). When it is 0, report "none,
 no change needed" and add no step. Only then inspect its optional
 `ignore_columns` field and submit; `--dry-run` reports `predicted_impact` or
-fails `no_op`. For a join, verify both schemas and expected multiplicity:
+fails `no_op`. A rule that removes nothing today but the user wants kept for future data
+("keep only rows where region is not empty") takes `--standing`; ask for it only on that intent. For a join, verify both schemas and expected multiplicity:
 
 ```bash
 mammoth schema get view.transform.discard-duplicates