@eigenpal/cli 0.20.26 → 0.20.28

package/dist/skill/SKILL.md

@@ -115,6 +115,23 @@ Use stable ids in scripts:
 - Workflow dataset example id: `evx_...`
 - Agent dataset example id: `aeg_...`
 
+## Document Parsing
+
+Use `ai.parse-v2` for new workflows. Start with only `input`: Eigenpal preserves native PDF text and recovers scanned/mixed pages through available, allowed OCR then vision. Text/Office parsing stays local. Do not add workflow-specific fallback branches.
+
+
+For mixed spreadsheet date/identifier columns, use `ai.parse-v2` with `output.includeCellMetadata: true`, or `transform.xlsx-to-json` with `valueMode: displayed` and `includeCellMetadata: true`. V2 includes raw/displayed value, type, address, number format, workbook date system and optional date rendering in page evidence and text. JSON conversion puts evidence in `cellMetadata`, linked by `outputRowIndex` and `key`; pass the full output to the extraction LLM, not just `rows`. Recorded date formats control rendering; headers and numeric magnitude do not classify dates. Preserve ambiguous identifiers/strings and ask extraction to flag conflicting evidence. Never guess missing date formats or clean names during parsing. Conversion trims only implicit empty trailing extent; explicit ranges and real distant cells retain limits. Formulas use cached values and are never evaluated.
+
+Before authoring, check the live server supports v2 with `eigenpal workflow step-type get ai.parse-v2`, then read `eigenpal docs read steps/ai/parse-v2`. Inspect `eigenpal models parser-readiness --json`; catalog readiness is not a live provider or worker-binary probe. Test native, scanned, and mixed inputs.
+
+Figure enrichment is independently opt-in; it may call vision even with native-only text transcription. Leave enrichment disabled for fully local processing.
+
+Configure `policy.native` (`prefer`, `require`, `skip`) and `policy.imageReading.order` only to express intentional restrictions/preferences. `providers.ocr` and `providers.vision` are exact pins. Deployment parsing defaults and provider boundaries apply; general text-generation defaults do not select the parser vision model.
+
+Use `output.textFormat` for representation and `output.require` for required evidence. Vision cannot supply word coordinates. Native spatial whitespace requires plain text. Unresolved pages fail by default; if the user explicitly chooses `advanced.allowPartial`, inspect `completeness` before downstream extraction.
+
+Legacy `ai.parse` remains supported. Never mix its flat `parseMode`/`nativeText` settings into v2. Migrate to a new file with `eigenpal workflow migrate-parser workflow.yaml --policy preserve --out workflow-v2.yaml` or `--policy auto` to adopt recommended defaults. Preserve mode retains OCR-only restrictions; automatic mode can broaden egress and costs. Review reported changes and representative outputs before publishing.
+
 ## Workflow Loop
 
 ```bash

package/dist/skill/reference/cli/models.md

@@ -8,12 +8,14 @@ Inspect models and providers configured for the current tenant environment (text
 - [Commands](#commands)
   - [Core](#core)
 - [Details](#details)
+  - [`eigenpal models parser-readiness [options]`](#eigenpal-models-parser-readiness-options)
   - [`eigenpal models list|ls [options]`](#eigenpal-models-listls-options)
 
 ## Surface
 
 ```
 models
+├── parser-readiness
 └── list|ls
 ```
 
@@ -21,12 +23,24 @@ models
 
 ### Core
 
-| Command                              | Description                                                                                                                                                                                                                       |
-| ------------------------------------ | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
-| `eigenpal models list\|ls [options]` | List configured text, vision, and OCR models. This is a catalog inventory from the server, not a live provider health probe. `health` is `configured` or `unconfigured` from local credentials. Pair with `--json` for scripting. |
+| Command                                      | Description                                                                                                                                                                                                                       |
+| -------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| `eigenpal models parser-readiness [options]` | Inspect deployment-approved ai.parse-v2 defaults. Catalog state only; document smoke tests are required for live readiness.                                                                                                       |
+| `eigenpal models list\|ls [options]`         | List configured text, vision, and OCR models. This is a catalog inventory from the server, not a live provider health probe. `health` is `configured` or `unconfigured` from local credentials. Pair with `--json` for scripting. |
 
 ## Details
 
+### `eigenpal models parser-readiness [options]`
+
+Inspect deployment-approved ai.parse-v2 defaults. Catalog state only; document smoke tests are required for live readiness.
+
+### Options
+
+| Flag               | Required | Default | Description                          |
+| ------------------ | -------- | ------- | ------------------------------------ |
+| `--base-url <url>` | no       |         | Server base URL                      |
+| `--json`           | no       |         | Emit machine-readable JSON on stdout |
+
 ### `eigenpal models list|ls [options]`
 
 List configured text, vision, and OCR models. This is a catalog inventory from the server, not a live provider health probe. `health` is `configured` or `unconfigured` from local credentials. Pair with `--json` for scripting.

package/dist/skill/reference/cli/surface.md

@@ -20,6 +20,7 @@ eigenpal
 │   ├── push
 │   ├── move <workflow-id>
 │   ├── delete <workflow-id>
+│   ├── migrate-parser <file>
 │   ├── folders
 │   │   ├── list|ls
 │   │   ├── create <path>
@@ -139,6 +140,7 @@ eigenpal
 │   └── env
 │       └── pull [target]
 ├── models
+│   ├── parser-readiness
 │   └── list|ls
 ├── email-servers|email-server
 │   ├── list|ls

package/dist/skill/reference/cli/workflow.md

@@ -23,6 +23,7 @@ Manage workflows: push, pull, and evaluate.
   - [`eigenpal workflow push [options]`](#eigenpal-workflow-push-options)
   - [`eigenpal workflow move [options] <workflow-id>`](#eigenpal-workflow-move-options-workflow-id)
   - [`eigenpal workflow delete [options] <workflow-id>`](#eigenpal-workflow-delete-options-workflow-id)
+  - [`eigenpal workflow migrate-parser [options] <file>`](#eigenpal-workflow-migrate-parser-options-file)
   - [`eigenpal workflow validate [options] [path]`](#eigenpal-workflow-validate-options-path)
   - [`eigenpal workflow clear-local [options] [examples...]`](#eigenpal-workflow-clear-local-options-examples)
   - [`eigenpal workflow folders list|ls [options]`](#eigenpal-workflow-folders-listls-options)
@@ -82,6 +83,7 @@ workflow
 ├── push
 ├── move <workflow-id>
 ├── delete <workflow-id>
+├── migrate-parser <file>
 ├── folders
 │   ├── list|ls
 │   ├── create <path>
@@ -155,6 +157,7 @@ workflow
 | `eigenpal workflow push [options]`                      | Create or update a workflow from a YAML file.                                                                                                                                                                                                                                                                                                                                                                                             |
 | `eigenpal workflow move [options] <workflow-id>`        | Move a workflow to a folder path, creating folders as needed                                                                                                                                                                                                                                                                                                                                                                              |
 | `eigenpal workflow delete [options] <workflow-id>`      | Delete a workflow definition while preserving past runs.                                                                                                                                                                                                                                                                                                                                                                                  |
+| `eigenpal workflow migrate-parser [options] <file>`     | Write a reviewable ai.parse-v2 workflow migration to a new file. Never pushes or changes the source.                                                                                                                                                                                                                                                                                                                                      |
 | `eigenpal workflow validate [options] [path]`           | Local-only validation. Without [path]: runs the templated three-way check (./workflow.yaml + ./evaluators.yaml + ./dataset/) in the project root. When the root has no workflow.yaml, discovers nested projects under eigenpal/workflows/<slug>/ or workflows/<slug>/ and validates each. With [path] pointing at a YAML file: validates that workflow.yaml only. For per-noun targeting use `evaluators validate` or `dataset validate`. |
 | `eigenpal workflow clear-local [options] [examples...]` | Delete local execution artifacts under ./dataset/examples/. Keeps the latest run per example by default.                                                                                                                                                                                                                                                                                                                                  |
 
@@ -352,6 +355,24 @@ Delete a workflow definition while preserving past runs.
 | `--base-url <url>` | no       |         | Server base URL                                                             |
 | `--json`           | no       |         | Emit machine-readable JSON on stdout                                        |
 
+### `eigenpal workflow migrate-parser [options] <file>`
+
+Write a reviewable ai.parse-v2 workflow migration to a new file. Never pushes or changes the source.
+
+### Arguments
+
+| Name   | Required | Variadic | Description |
+| ------ | -------- | -------- | ----------- |
+| `file` | yes      | no       |             |
+
+### Options
+
+| Flag                | Required | Default | Description                                                                         |
+| ------------------- | -------- | ------- | ----------------------------------------------------------------------------------- |
+| `--policy <policy>` | yes      |         | preserve backend restrictions, or auto to adopt automatic native/OCR/vision parsing |
+| `--out <file>`      | yes      |         | New YAML output file (must not already exist)                                       |
+| `--json`            | no       |         | Emit migration summary as JSON                                                      |
+
 ### `eigenpal workflow validate [options] [path]`
 
 Local-only validation. Without [path]: runs the templated three-way check (./workflow.yaml + ./evaluators.yaml + ./dataset/) in the project root. When the root has no workflow.yaml, discovers nested projects under eigenpal/workflows/<slug>/ or workflows/<slug>/ and validates each. With [path] pointing at a YAML file: validates that workflow.yaml only. For per-noun targeting use `evaluators validate` or `dataset validate`.