Fix Skill vs Task tool guidance in delegation-enforcement.md

Corrected the Core Rule that was causing 'Agent type not found' errors:
- Skills always use Skill tool (loads SKILL.md into main agent context)
- Agents/subagents always use Task tool (spawns isolated subprocess)
- Removed false claim that Skill tool 'replaces execution context'
- Clarified why skills that spawn subagents MUST use Skill tool
- Documented companion agent pattern as special case for non-subagent skills only
- Updated docs-operator description in CLAUDE.md accordingly

Author: mkaluzny

.claude-plugin/marketplace.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://anthropic.com/claude-code/marketplace.schema.json",
   "name": "maister-plugins",
-  "version": "2.0.0",
+  "version": "2.0.1",
   "description": "Structured, standards-aware development workflows for Claude Code",
   "owner": {
     "name": "Skillpanel",

plugins/maister-copilot/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "maister-copilot",
-  "version": "2.0.0",
+  "version": "2.0.1",
   "description": "Structured, standards-aware development workflows for Claude Code",
   "author": {
     "name": "Skillpanel",

plugins/maister-copilot/CLAUDE.md

@@ -717,7 +717,7 @@ Subagents are specialized AI agents invoked by skills and orchestrators. All age
 | Agent | Purpose | Invoked By | Details |
 |-------|---------|------------|---------|
 | `project-analyzer` | Deep codebase analysis for tech stack, architecture, conventions | `/maister-init` | `agents/project-analyzer.md` |
-| `docs-operator` | Internal service agent: executes docs-manager operations mid-workflow and returns control to the caller. Has docs-manager skill preloaded. | init, standards-update, standards-discover | `agents/docs-operator.md` |
+| `docs-operator` | Internal service agent: executes docs-manager operations mid-workflow via Task tool. Has docs-manager skill preloaded. **Special case**: companion agent pattern only works here because docs-manager does NOT spawn subagents (only file operations). Do not use this pattern for skills that spawn subagents. | init, standards-update, standards-discover | `agents/docs-operator.md` |
 | `task-classifier` | Classifies task descriptions into workflow types with confidence scoring | `/work` command | `agents/task-classifier.md` |
 | `gap-analyzer` | Compares current vs desired state with characteristic-detection-based analysis modules | development-orchestrator | `agents/gap-analyzer.md` |
 | `specification-creator` | Creates specs from gathered requirements with reusability search and self-verification | development-orchestrator, migration-orchestrator | `agents/specification-creator.md` |

plugins/maister-copilot/skills/orchestrator-framework/references/delegation-enforcement.md

@@ -6,12 +6,16 @@ Orchestrators must delegate work to skills and subagents instead of executing in
 
 **Always use Skill/Task tools to delegate. Never execute delegated work inline.**
 
-When a phase requires a skill or subagent:
-1. Use the `Skill` tool for skills **that are the last step in a phase/workflow** (Skill tool does not return control to the caller)
-2. Use the `Task` tool for subagents **and for mid-workflow skill operations** (Task tool returns control after completion)
+When a phase requires delegation:
+1. Use the **Skill tool** for **skills** — loads SKILL.md instructions into the main agent's context; the main agent executes the skill's instructions and continues with the orchestrator workflow afterward
+2. Use the **Task tool** for **subagents/agents** — spawns an isolated subprocess that returns results when complete
 3. Wait for completion before continuing
 
-**Important**: The Skill tool replaces the active execution context. Use it only when nothing needs to execute after it in the same workflow. For internal service skills that are invoked mid-workflow (like `docs-manager`), create a companion agent that preloads the skill via the `skills` frontmatter field and invoke it via Task tool instead.
+**Skills and agents are NOT interchangeable.** Skills always use Skill tool; agents always use Task tool. Never invoke a skill via Task tool (`subagent_type`) — it will fail with "Agent type not found."
+
+**Why skills MUST use Skill tool**: Skills like `codebase-analyzer`, `implementer`, and `implementation-verifier` spawn their own subagents (Explore agents, reporters, planners). Subagents cannot spawn other subagents — so these skills must run in the main agent context via Skill tool.
+
+**Companion agent pattern** (e.g., `docs-operator`): Only works for skills that do NOT spawn subagents (like `docs-manager` which only does file operations). A companion agent preloads the skill via the `skills` frontmatter field and is invoked via Task tool. This pattern fails for any skill that needs to spawn subagents.
 
 ---
 

plugins/maister/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "maister",
-  "version": "2.0.0",
+  "version": "2.0.1",
   "description": "Structured, standards-aware development workflows for Claude Code",
   "author": {
     "name": "Skillpanel",

plugins/maister/CLAUDE.md

@@ -717,7 +717,7 @@ Subagents are specialized AI agents invoked by skills and orchestrators. All age
 | Agent | Purpose | Invoked By | Details |
 |-------|---------|------------|---------|
 | `project-analyzer` | Deep codebase analysis for tech stack, architecture, conventions | `/maister:init` | `agents/project-analyzer.md` |
-| `docs-operator` | Internal service agent: executes docs-manager operations mid-workflow and returns control to the caller. Has docs-manager skill preloaded. | init, standards-update, standards-discover | `agents/docs-operator.md` |
+| `docs-operator` | Internal service agent: executes docs-manager operations mid-workflow via Task tool. Has docs-manager skill preloaded. **Special case**: companion agent pattern only works here because docs-manager does NOT spawn subagents (only file operations). Do not use this pattern for skills that spawn subagents. | init, standards-update, standards-discover | `agents/docs-operator.md` |
 | `task-classifier` | Classifies task descriptions into workflow types with confidence scoring | `/work` command | `agents/task-classifier.md` |
 | `gap-analyzer` | Compares current vs desired state with characteristic-detection-based analysis modules | development-orchestrator | `agents/gap-analyzer.md` |
 | `specification-creator` | Creates specs from gathered requirements with reusability search and self-verification | development-orchestrator, migration-orchestrator | `agents/specification-creator.md` |

plugins/maister/skills/orchestrator-framework/references/delegation-enforcement.md

@@ -6,12 +6,16 @@ Orchestrators must delegate work to skills and subagents instead of executing in
 
 **Always use Skill/Task tools to delegate. Never execute delegated work inline.**
 
-When a phase requires a skill or subagent:
-1. Use the `Skill` tool for skills **that are the last step in a phase/workflow** (Skill tool does not return control to the caller)
-2. Use the `Task` tool for subagents **and for mid-workflow skill operations** (Task tool returns control after completion)
+When a phase requires delegation:
+1. Use the **Skill tool** for **skills** — loads SKILL.md instructions into the main agent's context; the main agent executes the skill's instructions and continues with the orchestrator workflow afterward
+2. Use the **Task tool** for **subagents/agents** — spawns an isolated subprocess that returns results when complete
 3. Wait for completion before continuing
 
-**Important**: The Skill tool replaces the active execution context. Use it only when nothing needs to execute after it in the same workflow. For internal service skills that are invoked mid-workflow (like `docs-manager`), create a companion agent that preloads the skill via the `skills` frontmatter field and invoke it via Task tool instead.
+**Skills and agents are NOT interchangeable.** Skills always use Skill tool; agents always use Task tool. Never invoke a skill via Task tool (`subagent_type`) — it will fail with "Agent type not found."
+
+**Why skills MUST use Skill tool**: Skills like `codebase-analyzer`, `implementer`, and `implementation-verifier` spawn their own subagents (Explore agents, reporters, planners). Subagents cannot spawn other subagents — so these skills must run in the main agent context via Skill tool.
+
+**Companion agent pattern** (e.g., `docs-operator`): Only works for skills that do NOT spawn subagents (like `docs-manager` which only does file operations). A companion agent preloads the skill via the `skills` frontmatter field and is invoked via Task tool. This pattern fails for any skill that needs to spawn subagents.
 
 ---