feat: Erweitere Skill-Beschreibungen mit präziseren Nutzungsrichtlinien

- Verfeinere Description-Felder aller Skills für besseres Triggering - Neue skill-creator@claude-plugins-official Permission - /story mit neuer Reference-Struktur (phase-spezifische Detail-Dateien) - /go, /plan-review, /ship, /workflow-review: klarere Abgrenzungen - Vereinfachung im Vergleich zu vorigen Versionen Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-04-19 15:44:52 +02:00
parent 6f9079fa5f
commit 2fbb59852e
26 changed files with 1035 additions and 54 deletions
@@ -0,0 +1,81 @@
+Split: 13 train, 7 test (holdout=0.4)
+
+============================================================
+Iteration 1/5
+Description: Reviewt Skills, Agenten und Konfigurationsdateien dieses Repos auf Vollständigkeit, Konsistenz und Best Practices. Erstellt einen umsetzbaren Plan für Verbesserungen.
+============================================================
+Train: 24/39 correct, precision=100% recall=0% accuracy=62% (48.5s)
+  [FAIL] rate=0/3 expected=True: bitte prüfe alle skills und agenten auf konsistenz
+  [FAIL] rate=0/3 expected=True: schau mal über die workflow-konfiguration, besonders den /sh
+  [FAIL] rate=0/3 expected=True: reviewe den aktuellen workflow und sag mir was verbessert we
+  [FAIL] rate=0/3 expected=True: vergleiche unsere skills mit der externen best-practice-refe
+  [FAIL] rate=0/3 expected=True: sind die skills noch up to date?
+  [PASS] rate=0/3 expected=False: was bedeutet das feld context: fork in einem skill?
+  [PASS] rate=0/3 expected=False: wie wird der test-runner agent gestartet?
+  [PASS] rate=0/3 expected=False: reviewe meinen code in orders_service.py auf qualität
+  [PASS] rate=0/3 expected=False: bitte den security-audit für die neuen endpoints durchführen
+  [PASS] rate=0/3 expected=False: zeig mir alle hooks die aktuell konfiguriert sind
+  [PASS] rate=0/3 expected=False: kannst du das CLAUDE.md reviewen und verbessern?
+  [PASS] rate=0/3 expected=False: bitte den offenen PR #15 reviewen
+  [PASS] rate=0/3 expected=False: was sind die unterschiede zwischen /go und /ship?
+Test : 12/21 correct, precision=100% recall=0% accuracy=57% (0.0s)
+  [FAIL] rate=0/3 expected=True: ich möchte wissen ob meine claude-konfiguration best practic
+  [FAIL] rate=0/3 expected=True: gibt es widersprüche zwischen den skills in diesem repo?
+  [FAIL] rate=0/3 expected=True: workflow review bitte
+  [PASS] rate=0/3 expected=False: erstelle einen neuen skill für automatisches changelog-schre
+  [PASS] rate=0/3 expected=False: wie funktioniert der story-lifecycle?
+  [PASS] rate=0/3 expected=False: prüfe meinen implementierungsplan
+  [PASS] rate=0/3 expected=False: check mal ob settings.json korrekt ist
+
+Improving description...
+Traceback (most recent call last):
+  File "<frozen runpy>", line 198, in _run_module_as_main
+  File "<frozen runpy>", line 88, in _run_code
+  File "/home/martin/.claude/plugins/cache/claude-plugins-official/skill-creator/unknown/skills/skill-creator/scripts/run_loop.py", line 332, in <module>
+    main()
+    ~~~~^^
+  File "/home/martin/.claude/plugins/cache/claude-plugins-official/skill-creator/unknown/skills/skill-creator/scripts/run_loop.py", line 297, in main
+    output = run_loop(
+        eval_set=eval_set,
+    ...<11 lines>...
+        log_dir=log_dir,
+    )
+  File "/home/martin/.claude/plugins/cache/claude-plugins-official/skill-creator/unknown/skills/skill-creator/scripts/run_loop.py", line 202, in run_loop
+    new_description = improve_description(
+        client=client,
+    ...<7 lines>...
+        iteration=iteration,
+    )
+  File "/home/martin/.claude/plugins/cache/claude-plugins-official/skill-creator/unknown/skills/skill-creator/scripts/improve_description.py", line 114, in improve_description
+    response = client.messages.create(
+        model=model,
+    ...<5 lines>...
+        messages=[{"role": "user", "content": prompt}],
+    )
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_utils/_utils.py", line 283, in wrapper
+    return func(*args, **kwargs)
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/resources/messages/messages.py", line 1000, in create
+    return self._post(
+           ~~~~~~~~~~^
+        "/v1/messages",
+        ^^^^^^^^^^^^^^^
+    ...<30 lines>...
+        stream_cls=Stream[RawMessageStreamEvent],
+        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+    )
+    ^
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_base_client.py", line 1368, in post
+    return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))
+                           ~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_base_client.py", line 1062, in request
+    request = self._build_request(options, retries_taken=retries_taken)
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_base_client.py", line 521, in _build_request
+    headers = self._build_headers(options, retries_taken=retries_taken)
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_base_client.py", line 451, in _build_headers
+    self._validate_headers(headers_dict, custom_headers)
+    ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/martin/.local/lib/python3.13/site-packages/anthropic/_client.py", line 196, in _validate_headers
+    raise TypeError(
+        '"Could not resolve authentication method. Expected either api_key or auth_token to be set. Or for one of the `X-Api-Key` or `Authorization` headers to be explicitly omitted"'
+    )
+TypeError: "Could not resolve authentication method. Expected either api_key or auth_token to be set. Or for one of the `X-Api-Key` or `Authorization` headers to be explicitly omitted"