@aipermission/mcp 0.2.11 → 0.2.13

package/README.md

@@ -66,6 +66,12 @@ RabbitMQ, Docker, Kubernetes, and future connectors share the same model: target
 credential profile, connector action, token action permission, approval,
 history, and audit.
 
+Projects group connector targets for one local developer. Each MCP token has an
+enabled project scope in addition to its target/profile/action grants. Targets
+from disabled projects are omitted from discovery and rejected on direct calls;
+their saved action grants are preserved for later re-enablement. Projects are
+not multi-user or hosted RBAC.
+
 For SSH, call `get_connector_actions(target_ref)` to discover actions such as
 `exec`, `read_console`, `restart_console_session`, `browse_remote_files`, and
 `start_file_download`. SSH `exec` is intended for non-interactive commands. Use

package/dist/resources/aipermission-operator/SKILL.md

@@ -28,7 +28,9 @@ Before acting:
 5. Call `call_connector_action(target_ref, action_name, input, reason)`.
 
 If no target is visible, say that the current token has no accessible connector
-targets. If an action grant includes `expires_at`, treat access as temporary and
+targets. A target can be absent because its project is disabled for the token or
+because no effective action grant exists; do not claim that it was deleted or
+is offline. If an action grant includes `expires_at`, treat access as temporary and
 finish within that maintenance window or ask the operator to extend access.
 
 Target visibility is permission-scoped, not a live health check. A visible SSH
@@ -151,6 +153,30 @@ explicitly supports that behavior.
 MCP connector responses never include file contents, gateway temp paths, archive
 staging paths, or local upload contents.
 
+## S3/Object Storage Practice
+
+For S3 connector targets, discover actions with `get_connector_actions` and
+prefer this sequence:
+
+1. Use `bucket_info` only to verify bucket reachability or endpoint metadata.
+2. Use `list_objects` to browse. Pass `prefix` to enter a folder-like prefix,
+   and use `browse_input` from directory entries when it is returned.
+3. Use `cursor` from `next_cursor` or `next_page_input` for pagination. Do not
+   send fields named `continuation_token`; S3 pagination tokens are not
+   credentials here, but the connector exposes them as `cursor` to avoid
+   secret-like input names.
+4. Use `get_object_metadata` before `download_object` when you only need size,
+   type, headers, or existence.
+5. Use `download_object` only for bounded object reads explicitly requested by
+   the operator.
+6. Keep `overwrite=false` for `upload_object` and `rename_object` unless the
+   operator explicitly approved replacement.
+7. Treat `delete_object` as destructive and ask for explicit confirmation if
+   approval mode does not already provide it.
+
+Object content may contain secrets or customer data. Avoid downloading or
+echoing object contents unless the operator asked for that exact object.
+
 ## Secret Hygiene
 
 Command text, action input, action output, history, audit records, and console

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aipermission/mcp",
-  "version": "0.2.11",
+  "version": "0.2.13",
   "mcpName": "io.github.aipermission/aipermission-mcp",
   "description": "Local-first MCP bridge for the aipermission gateway.",
   "license": "AGPL-3.0-only",

package/server.json

@@ -3,12 +3,12 @@
   "name": "io.github.aipermission/aipermission-mcp",
   "title": "AIPermission",
   "description": "Local-first MCP bridge for the AIPermission gateway.",
-  "version": "0.2.11",
+  "version": "0.2.13",
   "packages": [
     {
       "registryType": "npm",
       "identifier": "@aipermission/mcp",
-      "version": "0.2.11",
+      "version": "0.2.13",
       "transport": {
         "type": "stdio"
       }