feature: retire azure-hosted-copilot-sdk skill

Author: JasonYeMSFT

.github/CODEOWNERS

@@ -20,7 +20,6 @@
 /plugin/skills/azure-deploy/ @tmeschter @wbreza @kvenkatrajan @paulyuk @RickWinter
 /plugin/skills/azure-diagnostics/ @tmeschter @saikoumudi @RickWinter
 /plugin/skills/azure-enterprise-infra-planner/ @Jbrocket @micha31r @arunrab @RickWinter
-/plugin/skills/azure-hosted-copilot-sdk/ @tmeschter @JasonYeMSFT @RickWinter
 /plugin/skills/azure-kubernetes/ @saikoumudi @chandraneel @gambtho @RickWinter
 /plugin/skills/azure-kusto/ @saikoumudi @RickWinter
 /plugin/skills/azure-messaging/ @kashifkhan @RickWinter

.github/skills/sensei/SKILL.md

@@ -178,7 +178,7 @@ metadata:
 
 > ⚠️ **"DO NOT USE FOR:" carries context-dependent risk.** In multi-skill environments (10+ skills with overlapping domains), anti-trigger clauses introduce the very keywords that cause wrong-skill activation on Claude Sonnet and fast-pattern-matching models ([evidence](https://gist.github.com/kvenkatrajan/52e6e77f5560ca30640490b4cc65d109)). For small, isolated skill sets (1-5 skills), the risk is low. When in doubt, use positive routing with `WHEN:` and distinctive quoted phrases.
 >
-> **Exception:** `DO NOT USE FOR:` is **REQUIRED** when a specialized skill's triggers overlap with a broader skill (e.g., `azure-hosted-copilot-sdk` vs. `azure-prepare` on "deploy to Azure"). Without the negative discriminator, the broader skill captures prompts that should route to the specialized one. Always run integration tests before removing a `DO NOT USE FOR:` clause.
+> **Exception:** `DO NOT USE FOR:` is **REQUIRED** when a specialized skill's triggers overlap with a broader skill (e.g., `azure-prepare` on "deploy to Azure"). Without the negative discriminator, the broader skill captures prompts that should route to the specialized one. Always run integration tests before removing a `DO NOT USE FOR:` clause.
 
 ## Test Scaffolding
 

.github/skills/sensei/references/SCORING.md

@@ -166,7 +166,7 @@ Per the [agentskills.io spec](https://agentskills.io/specification), the `name`
 | Single skill or small set (1-5 skills) with clear domain boundaries | Low | Anti-triggers are low-risk — domain boundaries are obvious |
 | Medium skill set (5-15 skills) with some overlap | Moderate | Anti-trigger keywords start competing with other skills' triggers |
 | Large skill set (15+ skills) with overlapping domains | **High** | Keyword contamination is measurable — negative keywords become activation keywords on fast-pattern-matching models |
-| **Specialized skill with trigger overlap against a broader skill** | **REQUIRED** | Anti-triggers are the **only** disambiguation mechanism when a specialized skill (e.g., `azure-hosted-copilot-sdk`) competes with a broader skill (e.g., `azure-prepare`) on shared trigger phrases like "deploy to Azure". Removing them causes routing regressions. |
+| **Specialized skill with trigger overlap against a broader skill** | **REQUIRED** | Anti-triggers are the **only** disambiguation mechanism when a specialized skill competes with a broader skill (e.g., `azure-prepare`) on shared trigger phrases like "deploy to Azure". Removing them causes routing regressions. |
 
 **Why large skill sets are risky:** On Claude Sonnet and similar models that use fast pattern matching (first ~20 words), `DO NOT USE FOR: Function apps` causes Sonnet to key on "Function apps" and **activate** the skill for Functions queries. This was empirically demonstrated across 24 Azure skills ([analysis](https://gist.github.com/kvenkatrajan/52e6e77f5560ca30640490b4cc65d109)). Anthropic's own published skills confirm this pattern — 4 of 5 skills in `anthropics/skills` use positive-only routing.
 
@@ -327,7 +327,6 @@ function scoreSkill(skill):
 
     # Warn on anti-triggers (keyword contamination risk)
     # Exception: DO NOT warn if skill has trigger overlap with a broader skill
-    # (e.g., azure-hosted-copilot-sdk overlaps with azure-prepare on "deploy")
     # In that case, DO NOT USE FOR is REQUIRED for disambiguation
     if containsAntiTriggers(skill.description):
         if hasOverlappingTriggersWithBroaderSkill(skill):

.github/skills/skill-authoring/references/guidelines/frontmatter.md

@@ -56,4 +56,4 @@ description: "Helps with Azure stuff."
 
 > ⚠️ **Do NOT add "DO NOT USE FOR:" clauses** unless the skill has trigger overlap with a broader skill (e.g., `azure-prepare`). For most skills, `DO NOT USE FOR:` causes keyword contamination on Claude Sonnet and similar models — use positive routing with distinctive `WHEN:` phrases instead.
 >
-> **Exception — disambiguation-critical skills:** When a specialized skill shares trigger phrases with a broader skill (e.g., both `azure-hosted-copilot-sdk` and `azure-prepare` match "deploy to Azure"), `DO NOT USE FOR:` is **REQUIRED** to prevent the broader skill from capturing the specialized skill's prompts. Removing it causes routing regressions. Always run integration tests before removing a `DO NOT USE FOR:` clause.
+> **Exception — disambiguation-critical skills:** When a specialized skill shares trigger phrases with a broader skill, `DO NOT USE FOR:` is **REQUIRED** to prevent the broader skill from capturing the specialized skill's prompts. Removing it causes routing regressions. Always run integration tests before removing a `DO NOT USE FOR:` clause.