feat: Update eval file paths in documentation and scripts to use workspace structure for improved organization

Author: aalmada

.github/skills/copilot-skill-creator/SKILL.md

@@ -366,7 +366,7 @@ Present the eval set to the user for review using the HTML template:
    - `__EVAL_DATA_PLACEHOLDER__` → the JSON array of eval items (no quotes around it — it's a JS variable assignment)
    - `__SKILL_NAME_PLACEHOLDER__` → the skill's name
    - `__SKILL_DESCRIPTION_PLACEHOLDER__` → the skill's current description
-3. Write to a temp file (e.g., `/tmp/eval_review_<skill-name>.html`) and open it: `open /tmp/eval_review_<skill-name>.html`
+3. Write to the skill's workspace (e.g., `<skill-name>-workspace/eval_review_<skill-name>.html`) and open it: `open <skill-name>-workspace/eval_review_<skill-name>.html`
 4. The user can edit queries, toggle should-trigger, add/remove entries, then click "Export Eval Set"
 5. The file downloads to `~/Downloads/eval_set.json` — check the Downloads folder for the most recent version in case there are multiple (e.g., `eval_set (1).json`)
 
@@ -437,8 +437,8 @@ When using an AI client without subagent support, the core workflow is the same
 
 **Updating an existing skill**: The user might be asking you to update an existing skill, not create a new one. In this case:
 - **Preserve the original name.** Note the skill's directory name and `name` frontmatter field -- use them unchanged. E.g., if the installed skill is `research-helper`, output `research-helper.skill` (not `research-helper-v2`).
-- **Copy to a writeable location before editing.** The installed skill path may be read-only. Copy to `/tmp/skill-name/`, edit there, and package from the copy.
-- **If packaging manually, stage in `/tmp/` first**, then copy to the output directory -- direct writes may fail due to permissions.
+- **Copy to a writeable location before editing.** The installed skill path may be read-only. Copy to `<skill-name>-workspace/skill-copy/`, edit there, and package from the copy.
+- **If packaging manually, stage in the workspace directory first** (e.g., `<skill-name>-workspace/`), then copy to the output directory -- direct writes may fail due to permissions.
 
 ---
 

.github/skills/copilot-skill-creator/scripts/run_eval.py

@@ -22,9 +22,7 @@
 def find_project_root() -> Path:
     """Find the project root by walking up from cwd looking for .github/.
 
-    Ensures the temporary skill file created during eval lands under the
-    project's .github/skills/ directory, which is where the Copilot CLI
-    looks for skill definitions.
+    Used to construct the workspace path where eval artifacts are stored.
     """
     current = Path.cwd()
     for parent in [current, *current.parents]:
@@ -46,17 +44,17 @@ async def run_single_query_async(
 
     Injects the skill into the session via system_message and uses
     on_pre_tool_use to detect — and immediately deny — any read_file call
-    targeting the skill.  No files are written to the workspace, avoiding
-    VS Code file-watcher churn when many queries run in parallel.
-    The ``project_root`` parameter is retained for API compatibility but
-    is no longer used.
+    targeting the skill.  No files are written to disk — the path is a
+    fictional reference used purely as a marker string; detection fires in
+    on_pre_tool_use before the read executes and the call is denied immediately.
     """
     unique_id = uuid.uuid4().hex[:8]
     clean_name = f"{skill_name}-skill-{unique_id}"
-    # Point to /tmp so no workspace files are touched (VS Code watches .github/).
-    # The file does not need to exist: detection fires in on_pre_tool_use before
-    # the read executes, and we deny the call immediately after.
-    skill_file_ref = f"/tmp/copilot-skill-evals/{clean_name}/SKILL.md"
+    # Point to the skill's workspace folder so all eval artifacts stay in one
+    # place alongside other iteration outputs. The file does not need to exist:
+    # detection fires in on_pre_tool_use before the read executes, and we deny
+    # the call immediately after — nothing is ever written to disk.
+    skill_file_ref = f"{project_root}/.github/skills/{skill_name}-workspace/copilot-skill-evals/{clean_name}/SKILL.md"
     system_message = {
         "content": (
             "You have access to the following skills. When a user's request matches "