docs(sdk): clarify pluginRoot scope for runtime plugins — #285 (activation-context: ActivateContext n'expose ni répertoire d'installation ni racine package ni file URL; commands-and-feedback + services: runCommand project/workspace-scoped, aucune résolution package-relative des scripts packagés; manifest: substitution ${pluginRoot}/${appDataDir} étendue par l'hôte uniquement dans contributes.mcpServers command/args/env/cwd; QA verte: npm run build, cohérence champs vs types SDK vérifiée)

docs/commands-and-feedback.md

@@ -43,6 +43,9 @@ Check preconditions before calling `runCommand()`:
 - `ownerAgentId` is a real agent id when the work is owned by an agent workflow.
   Do not use all-zero placeholders in production.
 - `cwd` is relative to the project root.
+- `command`, `args` and `cwd` do not resolve package-relative plugin resources.
+  Do not use `runCommand()` to launch scripts shipped inside the plugin package;
+  there is no public runtime `pluginRoot` path in `activate(ctx)`.
 - `command` and `args` are separate values. Do not shell-join user input.
 
 If any required precondition fails, return a skipped result:
@@ -196,3 +199,20 @@ await ctx.services?.tasks.runCommand({
 
 This creates misleading Work history, uses a placeholder owner and turns a
 precondition failure into a fake task. Return `status: "skipped"` instead.
+
+Do not try to reach packaged plugin scripts through workspace-relative paths:
+
+```ts
+await ctx.services?.tasks.runCommand({
+  projectId: project.id,
+  ownerAgentId,
+  label: "Run packaged tests",
+  command: "scripts/run-unity-tests.sh",
+  cwd: "."
+});
+```
+
+`scripts/run-unity-tests.sh` above is resolved like any other executable visible
+from the project task environment; it is not resolved relative to the plugin
+package. Use a project-owned script or an executable available on `PATH` until a
+dedicated packaged-resource API exists.