ellf-cli 4.0.96__tar.gz → 4.0.97__tar.gz

{ellf_cli-4.0.96/ellf_cli.egg-info → ellf_cli-4.0.97}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ellf-cli
-Version: 4.0.96
+Version: 4.0.97
 Summary: Ellf Command Line Interface
 Home-page: https://prodi.gy
 Author: ExplosionAI GmbH

{ellf_cli-4.0.96 → ellf_cli-4.0.97}/ellf_cli/about.json

@@ -1,7 +1,7 @@
 {
   "title": "Ellf CLI",
   "name": "ellf-cli",
-  "version": "4.0.96",
+  "version": "4.0.97",
   "summary": "Ellf Command Line Interface",
   "uri": "https://prodi.gy",
   "prog": "ellf",

{ellf_cli-4.0.96 → ellf_cli-4.0.97}/ellf_cli/ellf.json

@@ -1,7 +1,7 @@
 {
   "prog": "ellf",
   "help": "Ellf Command Line Interface.",
-  "version": "4.0.95",
+  "version": "4.0.96",
   "extra_key": "_extra",
   "commands": {
     "actions": {

{ellf_cli-4.0.96 → ellf_cli-4.0.97}/ellf_cli/ellf_skills/skills/ellf-monitor.assistant/SKILL.md

@@ -75,19 +75,29 @@ If follow-up action is needed, route to `/ellf-ops` or `/ellf-handoff`.
 
 ### Annotation tasks
 
-Use cluster calls such as:
+Use cluster calls for job status and per-dataset counts; use PAM for dataset
+discovery:
 ```text
 mcp__cluster__cluster_request(method="GET", path="/api/v1/jobs/{id}/status")
-mcp__cluster__cluster_request(method="GET", path="/api/v1/datasets/example-count", query={"dataset": "<name>"})
+mcp__pam__dataset_list(cluster_id="<id>")               # discover datasets (org-wide, registered with PAM)
+mcp__cluster__dataset_example_count(name="<name>")      # total examples in the dataset (per-cluster, via broker)
+mcp__cluster__dataset_session_counts(name="<name>")     # per-annotator breakdown (keys are session_ids; null = bulk import)
 ```
 
+`dataset_list` is a PAM read (org-wide registry), while the count tools talk to
+the broker because example counts live in the per-cluster Prodigy database.
+All three are read-only and do not require user confirmation. They do not
+read annotation contents — only counts and metadata.
+
 Read when needed:
 - `${CLAUDE_SKILL_DIR}/references/annotation_metrics.md`
 
 Report:
 - task state
 - whether the task appears reachable and healthy
-- annotation count
+- annotation count (from `dataset_example_count`)
+- per-annotator activity (from `dataset_session_counts`) when the user asks
+  who annotated what or about active annotators
 - dataset growth
 - whether agent assignment appears to be producing data if applicable
 

{ellf_cli-4.0.96/ellf_cli/ellf_skills/skills/ellf-monitor.coding → ellf_cli-4.0.97/ellf_cli/ellf_skills/skills/ellf-monitor.assistant}/references/annotation_metrics.md

@@ -4,10 +4,13 @@ Use this reference to interpret annotation activity and quality signals.
 
 ## What to watch
 
-- Total annotation count
-- Dataset growth over time
-- Number of active annotators or agents
-- Whether a running task is producing new examples
+- Total annotation count — from `mcp__cluster__dataset_example_count`
+- Dataset growth over time — compare counts across checks
+- Number of active annotators or agents — from `mcp__cluster__dataset_session_counts`
+  (each non-null key is a session_id; `null` means examples written directly via
+  `db-in` rather than through an annotation session)
+- Whether a running task is producing new examples — combine job status with
+  example count over time
 
 ## Warning signals
 

{ellf_cli-4.0.96 → ellf_cli-4.0.97/ellf_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ellf-cli
-Version: 4.0.96
+Version: 4.0.97
 Summary: Ellf Command Line Interface
 Home-page: https://prodi.gy
 Author: ExplosionAI GmbH