@su-record/vibe 3.2.1 → 3.2.3

package/skills/vibe.educational-content/references/learning-design.md

@@ -0,0 +1,81 @@
+# Learning design
+
+Use this reference to align objectives, activities, feedback, and assessment.
+
+## Write observable objectives
+
+Use `learner + action + conditions + criterion` when the criterion matters.
+
+- weak: understand photosynthesis
+- observable: explain how light energy becomes stored chemical energy using a labeled diagram
+- weak: know Git branches
+- observable: create a feature branch, resolve one merge conflict, and explain why each resolution is correct
+
+Match the action to the desired depth:
+
+- remember: identify, recall, label
+- explain: summarize, compare, classify, explain
+- apply: use, calculate, execute, demonstrate
+- analyze: diagnose, distinguish, map, justify
+- create: design, produce, compose, implement
+- evaluate: critique, prioritize, defend with criteria
+
+## Align the learning sequence
+
+For every objective, provide:
+
+1. a model or example of the target performance
+2. guided practice that exposes learner thinking
+3. feedback specific to the success criteria
+4. independent practice without answer leakage
+5. an assessment using the same skill at an appropriate difficulty
+6. a transfer task in a changed context when durable learning matters
+
+Do not assess a deeper behavior than the instruction and practice prepared. Do not use recall questions as proof of application or analysis.
+
+## Diagnose the audience
+
+Distinguish:
+
+- prerequisite knowledge: facts or skills needed before starting
+- misconceptions: plausible but incorrect mental models
+- motivation: why the outcome matters to the learner
+- constraints: time, language, device, accessibility, instructor support
+
+Use a short pre-check when assumptions materially affect the path. Offer a bypass for learners who already demonstrate the prerequisite.
+
+## Scaffold and fade
+
+Move from worked example → completion task → independent task. Reduce hints as competence grows. Keep essential information near the activity that uses it. Split intrinsic complexity into meaningful chunks without fragmenting the overall mental model.
+
+## Give feedback
+
+Make feedback:
+
+- timely: close enough to the attempt to guide correction
+- specific: name the criterion and evidence
+- actionable: identify the next attempt
+- explanatory: address the learner's reasoning, not only correctness
+
+Avoid praise or correction that reveals no path forward. For high-stakes assessment, separate coaching practice from scored performance.
+
+## Design fair assessment
+
+Check:
+
+- validity: does the task measure the stated objective?
+- reliability: would equivalent performance receive equivalent judgment?
+- accessibility: can learners perceive and respond without irrelevant barriers?
+- fairness: does success depend on unstated cultural, linguistic, or technical knowledge?
+- authenticity: does the task resemble meaningful use of the skill?
+
+Provide acceptable variants and a rubric for open-ended work. Never fabricate learner results or claim effectiveness without evidence.
+
+## Adapt by format
+
+- self-study: make navigation, answers, and recovery self-contained
+- instructor-led: include facilitation prompts, timing, observation cues, and optional extensions
+- video: use retrieval pauses; do not replace practice with passive viewing
+- short-form: teach one distinction or action and give one immediate retrieval prompt
+- reference material: optimize lookup and decision support; do not force an unnecessary lesson arc
+- onboarding: use the real performance context, safe fixtures, and operational boundaries

package/skills/vibe.educational-content/references/output-patterns.md

@@ -0,0 +1,109 @@
+# Output patterns
+
+Select one primary pattern and omit unused fields.
+
+## Article or tutorial
+
+```markdown
+# Title: learner-visible outcome
+Audience:
+Prerequisites:
+Tested with:
+
+## What you will build
+## Mental model
+## Worked example
+## Step 1 — action
+Exact input:
+Expected result:
+Verification:
+## Common failure and recovery
+## Independent practice
+## Assessment and answer guidance
+## Extension
+## Recap
+## Sources and limitations
+```
+
+## Screencast or course chapter
+
+| Time | Learning beat | Narration intent | Visual action | Learner prompt | Evidence shot | Asset |
+|---|---|---|---|---|---|---|
+| 00:00 | Promise | Name the outcome | Show final state | Predict or notice | Working result | capture |
+
+Add a capture manifest:
+
+```yaml
+- id: proof-final
+  state: exact prerequisite state
+  action: command, prompt, or UI action
+  evidence: expected visible result
+  privacy: crop or redact requirements
+  retry: recovery action
+```
+
+## Lesson or workshop
+
+```markdown
+Outcome:
+Duration:
+Starting repository/state:
+Prior-knowledge check:
+
+Checkpoint 1:
+- learner action
+- facilitator proof
+- likely misconception
+- feedback or recovery
+
+Independent task:
+Acceptance criteria:
+Answer guidance or rubric:
+Cleanup or rollback:
+```
+
+## Short-form educational content
+
+```markdown
+Hook: surprising observable result
+Context: one sentence
+Teach: one model, contrast, or no more than three actions
+Retrieve: one prediction, decision, or recall prompt
+Proof: answer, demonstration, or comparison
+Caveat: scope, evidence, safety, or limitation
+CTA: reproducible next step, not engagement bait
+```
+
+## Assessment item
+
+```yaml
+objective: "observable learning outcome"
+prompt: "task that elicits the outcome"
+evidence: "what a valid response must demonstrate"
+criteria:
+  - name: "criterion"
+    levels: ["not yet", "developing", "meets", "extends"]
+answer_guidance: "reasoning and acceptable variants"
+misconceptions:
+  - signal: "common wrong pattern"
+    feedback: "specific corrective prompt"
+fairness_check: "language, accessibility, cultural, or prerequisite issue"
+```
+
+## Verification appendix
+
+```yaml
+environment:
+  context: ""
+  versions: []
+claims:
+  - claim: ""
+    status: verified | inferred | uncertain
+    evidence: ""
+    checked_at: "YYYY-MM-DD"
+checks:
+  - learner_action: ""
+    expected: ""
+    observed: ""
+unresolved: []
+```

package/skills/vibe.educational-content/references/source-method.md

@@ -0,0 +1,40 @@
+# Source-derived method
+
+Use this reference when adapting the method demonstrated in [Yalco's Codex training video](https://youtu.be/Bzdcg70ivrs). It records reusable principles, not a transcript.
+
+## Evidence basis
+
+- YouTube metadata identifies the video as “코덱스 참교육 - Codex 업무 자동화부터 바이브 코딩까지” by 얄팍한 코딩사전.
+- The creator's official course page identifies the public video as the 13:19 introduction, “코덱스 소개 & 강의에서 배울 것.” It describes the course examples and method across the Codex app, editor, CLI, plugins, automation, skills, MCP, testing, review, and harness engineering.
+- Full captions and frame-level playback were unavailable during source analysis because YouTube blocked the server request. Do not present the wider course curriculum as a scene-by-scene account of the 13:19 video.
+- Exact UI labels, model names, generated configuration, availability, and account requirements are version-sensitive. Verify them against current primary documentation before publishing.
+
+## Reusable principles
+
+1. Contrast a chat answer with an agent changing observable state.
+2. Start beginners on the lowest-friction surface, then reveal editor and CLI depth.
+3. Teach through a continuous result: create a file, inspect it, render it, then extend it.
+4. Show exact prompts or commands alongside their actual effects.
+5. Verify in the destination system: browser, filesystem, calendar, draft folder, or job history.
+6. Keep human approval around file deletion, messages, purchases, publication, and other external writes.
+7. Isolate agent work with a bounded project folder or sandbox.
+8. Explain that generated details may vary while acceptance criteria remain stable.
+9. Promote a successful one-off task to scheduled automation only after manual validation.
+10. Give automation an explicit scope, cadence, observable failure, and way to disable it.
+11. Structure agent requests with a goal, context, constraints, and observable completion criteria.
+12. Teach adjustment and verification instead of promising output identical to a recording.
+
+## Content adaptation
+
+Translate the method, not the examples:
+
+- replace the demonstrated product with the target technology
+- replace “app → editor → CLI” with “easy surface → inspectable implementation → power-user surface”
+- replace calendar or email proof with the target system's independent readback
+- replace feature enumeration with a single outcome that crosses two or three depth levels
+
+Attribute ideas when closely discussing the source. Do not reproduce long wording, full transcripts, or the source's visual identity.
+
+## Reliability notes
+
+Treat any detailed timeline produced by automated video analysis as an index to verify, not as authoritative evidence. Recheck timestamps and screen text before including them in published content. If direct playback or captions are unavailable, label the gap and avoid exact quotations.

package/skills/{event-comms → vibe.event-comms}/SKILL.md

@@ -1,5 +1,5 @@
 ---
-name: event-comms
+name: vibe.event-comms
 invocation: [auto]
 tier: standard
 description: "Event communication automation — SMS via Aligo, email via Gmail OAuth, SNS posts for LinkedIn/Threads. Handles templates, BCC rules, and confirmation protocol."

package/skills/{event-ops → vibe.event-ops}/SKILL.md

@@ -1,5 +1,5 @@
 ---
-name: event-ops
+name: vibe.event-ops
 invocation: [auto]
 tier: standard
 description: "Event operations — nametags, checklists, operational slides, Slido plans, image generation specs, settlement reports."

package/skills/{event-planning → vibe.event-planning}/SKILL.md

@@ -1,5 +1,5 @@
 ---
-name: event-planning
+name: vibe.event-planning
 invocation: [auto]
 tier: standard
 description: "Community event automation — D-Day timeline, community patterns, proactive task execution. Auto-activates for event planning, meetup/webinar/conference management."