Fix sandboxes skill description exceeding 1024-char limit

The sandboxes SKILL.md frontmatter description was 1386 chars, which
exceeds the 1024-char limit enforced by Copilot CLI. The skill was
installed on disk but silently rejected at load time with:

  skill_errors: ["…/SKILL.md: Skill description must be at most 1024 characters"]

Trimmed to 1007 chars while preserving the what/when/triggers structure
and all primary trigger phrases. Dropped phrases that overlapped with
kept ones (e.g. "azure container apps sandbox", "ACA sandbox",
"isolated VM", "sandbox lifecycle", "copy files to sandbox",
"sandbox volume", "sandbox labels", "sandbox apply yaml",
"egress deny", "sandbox inception", "computer use sandbox").

Bumped version to 0.0.2-beta so installed clients pick up the fix.

Fixes #1726

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Author: Copilot

plugin/skills/sandboxes/SKILL.md

@@ -2,29 +2,24 @@
 name: sandboxes
 description: |
   Sandboxes — hardware-isolated microVMs on Azure Container Apps for
-  running AI-generated code, coding agents, MCP servers, web apps, and
+  AI-generated code, coding agents, MCP servers, web apps, and
   ephemeral workloads. Snapshot/resume, scale-to-zero, sub-second
-  startup, deny-default egress. Driven by the `aca` CLI (auth delegates
-  to `az login`).
+  startup, deny-default egress. Driven by the `aca` CLI (auth via
+  `az login`).
 
-  Use when the user wants to: create or manage sandbox groups and
-  sandboxes; exec commands or open an interactive shell; read/write
-  files; expose ports; snapshot, stop, resume, commit to a disk, or
-  mount volumes; tighten egress; manage secrets, managed identity,
-  labels; apply YAML specs; or run scenarios like web apps, coding
-  agents, code interpreter, swarms, sandbox inception, computer-use,
-  MCP hosting, data processing, or developer workflows.
+  Use when the user wants to: create/manage sandbox groups and
+  sandboxes; exec or open a shell; read/write files; expose ports;
+  snapshot, stop, resume, commit to disk; mount volumes; tighten
+  egress; manage secrets, identity, labels; apply YAML; or run
+  scenarios like web apps, coding agents, code interpreter, swarms,
+  computer-use, or MCP hosting.
 
   Triggers: "create sandbox", "sandbox group", "aca cli", "aca
-  sandbox", "azure container apps sandbox", "ACA sandbox", "microVM",
-  "isolated VM", "run untrusted code", "exec in sandbox", "sandbox
-  shell", "copy files to sandbox", "sandbox port", "sandbox snapshot",
-  "commit sandbox to disk", "sandbox volume", "mount volume sandbox",
-  "suspend sandbox", "resume sandbox", "sandbox lifecycle",
-  "auto-suspend sandbox", "sandbox secret", "sandbox managed identity",
-  "sandbox labels", "sandbox apply yaml", "egress deny", "egress
-  allow-list", "code interpreter", "agent swarm", "sandbox inception",
-  "coding agent sandbox", "computer use sandbox", "host mcp".
+  sandbox", "microVM", "run untrusted code", "exec in sandbox",
+  "sandbox shell", "sandbox port", "sandbox snapshot", "commit sandbox
+  to disk", "mount volume sandbox", "suspend/resume sandbox", "sandbox
+  secret", "sandbox managed identity", "egress allow-list", "code
+  interpreter", "agent swarm", "coding agent sandbox", "host mcp".
 ---
 
 # Sandboxes

plugin/skills/sandboxes/version.json

@@ -1,4 +1,4 @@
 {
   "name": "sandboxes",
-  "version": "0.0.1-beta"
+  "version": "0.0.2-beta"
 }