Merge pull request #13 from apple-ouyang/codex/avoid-nested-skill-discovery

Keep Git source installs to one SKILL.md entrypoint

Author: yaojingang

…rt-demo/geo-content-brief-skill/SKILL.md …geo-content-brief-skill/SKILL.example.md.previews/yao-meta-skill-v2-report-demo/geo-content-brief-skill/SKILL.md renamed to .previews/yao-meta-skill-v2-report-demo/geo-content-brief-skill/SKILL.example.md .previews/yao-meta-skill-v2-report-demo/geo-content-brief-skill/SKILL.md renamed to .previews/yao-meta-skill-v2-report-demo/geo-content-brief-skill/SKILL.example.md

@@ -304,6 +304,7 @@ That active install writes to `~/.agents/skills/yao-meta-skill` and can make Cod
 Keep this repository focused on the meta-skill factory.
 
 - Put reusable factory examples in `examples/`.
+- Name embedded example entrypoints `SKILL.example.md`; reserve the exact `SKILL.md` filename for the installable root skill so recursive agent discovery does not activate examples or test fixtures.
 - Put reusable benchmark evidence, regression results, and release evidence in `reports/`.
 - Keep private analysis reports, customer-specific outputs, and one-off generated business skills outside this repository unless they are intentionally promoted into an example or regression fixture.
 - Place real generated skills as sibling skill directories under the local skill workspace, not as top-level folders inside `yao-meta-skill`.

README.md

@@ -9,13 +9,13 @@
     },
     {
       "name": "team-frontend-review",
-      "description_file": "examples/team-frontend-review/generated-skill/SKILL.md",
+      "description_file": "examples/team-frontend-review/generated-skill/SKILL.example.md",
       "semantic_config": "examples/team-frontend-review/optimization/semantic_config.json",
       "threshold": 0.48
     },
     {
       "name": "governed-incident-command",
-      "description_file": "examples/governed-incident-command/generated-skill/SKILL.md",
+      "description_file": "examples/governed-incident-command/generated-skill/SKILL.example.md",
       "semantic_config": "examples/governed-incident-command/optimization/semantic_config.json",
       "threshold": 0.48
     }

evals/confusion/route_cases.json

@@ -1,5 +1,7 @@
 # Examples
 
+Example packages use `SKILL.example.md` instead of `SKILL.md`. Git-based installers copy the whole root skill directory, and some agents recursively discover every nested `SKILL.md`; the non-discoverable filename keeps examples from becoming active skills after installation. Rename it to `SKILL.md` after copying an example into its own standalone skill directory.
+
 This directory shows complete examples from raw workflow input to final skill package shape.
 
 It is organized in three tiers: