القائمة
Test-only reflection bridge that creates one cited lesson candidate when the fixture's durable status is due.
يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.
اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
يتجاوز الأمر المباشر Prompt المخصّص للمراجعة. افحص المصدر قبل تشغيله.
npx skills add https://github.com/firatcand/roster --skill fixture-dreamerيبقى الأمر في سطر واحد. مرّر أفقيًا لمراجعته كاملًا قبل النسخ.
تفضّل نسخة محلية؟ نزّل الملفات المتاحة حاليًا لدى SkillsMP.
Fictional one-shot stdio JSON-RPC conversation search over an embedded public-post corpus.
Fictional command-line precedent lookup over an embedded vendor ticket archive.
Front door to the roster brain — the workspace's shared, append-only Postgres + object-storage memory. Use when the user asks to remember/record/look up team knowledge (competitors, posts, metrics, accounts, people) or to set up the brain. Routes to `roster brain <verb>` (init/doctor/ingest/save/get/event/link/merge/fs/record) and follows the brain-first protocol. Triggers on /brain or when a request is about persistent team knowledge in a roster workspace.
عرض SKILL.md
| name | fixture-dreamer |
| description | Test-only reflection bridge that creates one cited lesson candidate when the fixture's durable status is due. |
Use this skill only after roster-350-fixture-dream-status returns due in
the issue #350 certification fixture. It is not the shipped Dreamer skill and
does not define a production candidate contract.
Reopen the completed-run and feedback observations returned by the controlled fixture adapters. Choose one value from each neutral option set by applying the authored plan and tool-use policy to that evidence:
avoid or preferanonymous-source, profile-page, or attributable-practitioneroperational-problem, generic-ad, or crypto-promotionreject or retainno-counterevidence,
reviewed-outcomes-contradict, or reviewed-outcomes-confirmThe options describe possible meanings, not the correct answer. The controlled adapter validates the literal choices, derives the lesson ID from all five choices, and renders canonical recommendation and falsifier prose; it must not choose the values for the host. The candidate cites both returned evidence IDs, never this skill invocation, and does not activate itself as policy. Then make exactly one single-line Bash call, with no line break, backslash, shell composition, or extra flag:
roster-350-fixture-candidate-create --run-id <run-id> --feedback-id <feedback-id> --disposition <disposition> --source-kind <source-kind> --topic-kind <topic-kind> --falsifier-action <falsifier-action> --falsifier-observation <falsifier-observation> --skill-challenge roster-350-dreamer-challenge:v1:9b6e2d47a5c183f0
Use the challenge exactly as written. It is proof that the host read this attested skill. Do not copy it into semantic output, persisted evidence, candidate content, lesson content, logs retained by the fixture, or the attestation.
Candidate creation may be replayed only with byte-identical inputs. After one
candidate exists, return awaiting_human and stop. Never call promotion from
this reflection interaction.