@fernado03/zoo-flow 0.7.4 → 0.7.6

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@fernado03/zoo-flow",
   "description": "Workflow control plane for Zoo Code.",
-  "version": "0.7.4",
+  "version": "0.7.6",
   "type": "module",
   "bin": {
     "zoo-flow": "bin/zoo-flow.js"

package/templates/full/.roo/rules-code-tweaker/01-completion.md

@@ -22,7 +22,7 @@ Before any `attempt_completion`, re-read the command body and confirm no later p
 
 The orchestrator forwards your output to the user. Include enough detail that the user can act on it without re-running commands.
 
-If the output is long (multi-file diff summary, full test output, verification report), write it to `.scratch/` first and reference the path in `attempt_completion`. The extension may truncate long completion messages, so files ensure nothing is lost.
+If the output is long, also write it to `.scratch/` for persistence and user reference. Mention the file path in `attempt_completion` so the user can open it. The orchestrator will not read the file — it only presents your completion message.
 
 Before running `git commit` or `git push`, halt and wait for explicit user approval. Never push unless explicitly asked.
 

package/templates/full/.roo/rules-custom-orchestrator/00-routing.md

@@ -6,13 +6,14 @@ Router only. No reading implementation files, editing, shell, or answering imple
 
 After `attempt_completion` from a worker, the orchestrator must:
 
-1. If the completion message references a `.scratch/` file, read and present that file's content.
-2. Show the worker's full output to the user without truncating or summarizing.
-3. Add one line: recommended next command (if any).
-4. Stop and wait for user input.
+1. Show the worker's full output to the user without truncating or summarizing.
+2. Add one line: recommended next command (if any).
+3. Stop and wait for user input.
 
 Do not route a new user message until the previous subtask result is fully presented. If the user sends a new message before the result was shown, finish presenting the result first, then route the new message.
 
+The `.scratch/` file referenced in the completion is for user reference only — the orchestrator does not read it.
+
 ## Hard delegation boundary
 
 For free-form requests, the only valid `new_task` target slugs are:

package/templates/full/.roo/rules-system-architect/02-completion.md

@@ -11,9 +11,9 @@ If you entered this window via `switch_mode` (you are mid-chain, not the entry p
 
 Before any `attempt_completion`, re-read the command body and confirm no later phase is assigned to another mode. If one is, `switch_mode` instead.
 
-Do not use `attempt_completion` to avoid required implementation work. Every `attempt_completion` must include the full output the user needs to see (review findings, exploration map, diagnosis, plan), not just summaries. The orchestrator will forward your output to the user; if you summarize, the user loses detail.
+Do not use `attempt_completion` to avoid required implementation work. Every `attempt_completion` must include the full output the user needs to see (review findings, exploration map, diagnosis, plan), not just summaries. The orchestrator forwards your output to the user; if you summarize, the user loses detail.
 
-If the output is long (multi-page review, full exploration map, detailed diagnosis), write it to `.scratch/` first and reference the path in `attempt_completion`. The extension may truncate long completion messages, so files ensure nothing is lost.
+If the output is long, also write it to `.scratch/` for persistence and user reference. Mention the file path in `attempt_completion` so the user can open it. The orchestrator will not read the file — it only presents your completion message.
 
 Halt for explicit user approval before testing a bug hypothesis, finalizing an architecture plan, publishing issues, or making irreversible decisions.