fix: break delegate→plan-mode→approve infinite loop

The Stop hook's continuation injection used `reason: "continue"` which the
model interpreted as a fresh work request. Combined with orchestrator_stub's
"all work MUST go via /workflow-orchestrator:delegate" rule and delegate.md's
unconditional EnterPlanMode at Stage 0, this caused an infinite
ExitPlanMode → Stop hook → continue → delegate → EnterPlanMode → ExitPlanMode
loop after every plan approval.

Three-layer fix:
- Stop hook continuation message is now explicit anti-redelegation
  (hooks/stop/python_stop_hook.py).
- Orchestrator stub adds an "Exception — continuation after plan approval"
  clause that overrides the routing rule during in-flight delegation
  (system-prompts/orchestrator_stub.md).
- delegate.md gets a "RE-INVOCATION GUARD" section as the first content
  after frontmatter, telling the skill to skip Stage 0 if it just exited
  plan mode (commands/delegate.md).

Layered guards because the loop is driven by model interpretation, not
deterministic state — any one layer alone could be missed by prompt cache
effects or session resumes; together they harden the path.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

Author: barkain

commands/delegate.md

@@ -4,6 +4,10 @@ argument-hint: [task description]
 allowed-tools: Agent, Task, EnterPlanMode, ExitPlanMode, AskUserQuestion, TaskCreate, TaskUpdate, TaskGet, TaskList, ToolSearch, TeamCreate, SendMessage
 ---
 
+## RE-INVOCATION GUARD (READ FIRST)
+
+Before doing ANYTHING else: if your most recent tool call in this conversation was `ExitPlanMode`, OR if you arrived here via a "PLAN ALREADY APPROVED" / "continuing to STAGE 1" continuation message, the plan is already approved. **Do NOT call `EnterPlanMode`. Do NOT re-enter Stage 0.** Skip directly to **STAGE 1: EXECUTION** — render the dependency graph from the approved plan in context and begin spawning Wave 0 agents. Re-entering plan mode here causes an infinite delegate→plan→approve loop.
+
 # Workflow Orchestrator System Prompt
 
 ## Purpose

hooks/stop/python_stop_hook.py

@@ -173,8 +173,13 @@ def check_workflow_continuation() -> bool:
         # This mimics ralph-wiggum's loop mechanism
         output = {
             "decision": "block",
-            "reason": "continue",
-            "systemMessage": "⚡ Workflow continuation: Proceeding to STAGE 1 execution.",
+            "reason": (
+                "PLAN ALREADY APPROVED. Execute Stage 1 NOW directly from the existing "
+                "approved plan in context. DO NOT call /workflow-orchestrator:delegate. "
+                "DO NOT call EnterPlanMode. DO NOT re-enter plan mode. "
+                "Render the dependency graph and start spawning Wave 0 agents."
+            ),
+            "systemMessage": "⚡ Continuing to STAGE 1 execution (plan already approved).",
         }
         print(json.dumps(output))  # noqa: T201
         logger.debug("Output block decision with 'continue' reason")

system-prompts/orchestrator_stub.md

@@ -10,6 +10,10 @@ Any user request that requires work — writing code, running tools, multi-step
 
 The main agent does not execute work tools directly. Use only: Tasks API, AskUserQuestion, and `/workflow-orchestrator:delegate`. The delegate command loads the full orchestrator (planning, agent assignment, execution waves) on demand.
 
+## Exception — continuation after plan approval
+
+If your most recent tool call was `ExitPlanMode`, OR you received a "PLAN ALREADY APPROVED" / "continuing to STAGE 1" continuation message, **do NOT re-invoke `/workflow-orchestrator:delegate`** and **do NOT call `EnterPlanMode`** again. The orchestrator is already loaded and the plan is already approved — proceed directly to Stage 1 execution by rendering the dependency graph and spawning Wave 0 agents. The "all work → delegate" rule above does NOT apply during in-flight delegation.
+
 ## Team Mode
 
 If `TeamCreate` is in your available tools, agent teams are enabled. When you run `/workflow-orchestrator:delegate`, default to team mode (`TeamCreate` + `Agent(team_name=...)`) for multi-agent work. If `TeamCreate` is not available, the delegate flow falls back to parallel subagents automatically.