원클릭으로 Manus에서 모든 스킬 실행

시작하기

runspace-pitch

스타0

포크0

업데이트2026년 5월 1일 00:29

Build / refine RunSpace 10-page deck, 90-second video, 3-minute English-subtitled video. Anonymity-enforced.

설치

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

Manus에서 실행

출처

thc1006

thc1006/OrbitOps-Copilot

GitHub 저장소 열기 Creator 저장소 보기

다운로드

Manus에서 실행

Overview

Translates technical artefacts into RunSpace-ready pitch material. All output is treated as public submission; absolute no-PII. Aligns to two scoring frames: 30/30/30/10 and 35/25/20/20 (docs/00_research_2026_04.md §5).

Triggers

Editing docs/06_runspace_pitch_outline.md, docs/07_demo_script_90s.md, docs/08_demo_script_3min.md.
Producing a slide deck (PDF) or video for submission.
Slash command /research produces a new fact that should be reflected on a slide.

Inputs

docs/01_product_strategy.md (UC1/UC2 + demo story).
docs/00_research_2026_04.md (sources + standards anchoring).
docs/03_breakthrough_directions.md (roadmap).
docs/10_links.md (allowlisted footnote URLs).

Step-by-step workflow

Open the matching outline / script doc.
Pull narrative beats from docs/01_product_strategy.md.
Anchor every external claim to a footnote citing a docs/10_links.md URL.
Cross-check against docs/00_research_2026_04.md confidence column — mark Medium/Low claims with hedge language.
Draft slide / scene; render in dry-run.
Anonymity sweep: scripts/check-no-secrets.sh over generated material.
(For final video / PDF) exiftool -all= <file> to strip metadata.
Run claims-audit skill against the new pitch text — every claim must classify cleanly.

Output format

Updated outline / script .md.
Optional tmp/runspace-deck.pdf and tmp/runspace-90s.mp4 (gitignored, but submitted externally).

Verification checklist

No team / school / personal identifiers in text, slides, video, or rendered metadata.
Every external claim has a footnote URL from docs/10_links.md.
Two-frame self-score updated if scoring narrative shifts.
./verify.sh clean (anonymity gate).
Video metadata stripped via exiftool -all=.
OS toolbar / browser tabs not visible in any screenshot.
claims-audit skill produced zero OVER-CLAIM rows.

Common failure modes

Slide footer auto-stamped with author identity from the slide tool (PowerPoint, Keynote).
Demo recording with terminal showing whoami / hostname in prompt.
A new claim added without a corresponding entry in docs/10_links.md.
"About us" slide accidentally drafted — must be removed entirely.

Forbidden actions

Naming team / school / individuals.
Quoting a metric / outcome that the implementation has not actually demonstrated (use claims-audit first).
Including any logo other than OSS project logos with attribution.
Using stock photography that has embedded metadata or watermarks.

이 저장소의 다른 Skills

같은 저장소

no-secrets-anonymity-check

thc1006/OrbitOps-Copilot

Run + interpret + extend the real-secrets scan. Pre-commit, pre-PR, pre-release gate. (Folder name retains "anonymity" for backwards-link compat; the anonymity-scan was retired 2026-05-01 — see CLAUDE.md §2.1.)

2026-05-010

claims-audit

thc1006/OrbitOps-Copilot

Audit every pitch / docs claim into one of {implemented, simulated, planned, external_reference}. Block over-claiming.

2026-05-010

demo-replay-hardening

thc1006/OrbitOps-Copilot

Make the AC-004 demo replay deterministic, recordable, and resilient to network blips / LLM provider failure.

2026-05-010

k8s-demo-deploy

thc1006/OrbitOps-Copilot

Build / iterate the kind / k3d / Kustomize / Helm demo deployment. Use for any change under deploy/ or packages/nephio-stubs/.

2026-05-010

llm-grounding-review

thc1006/OrbitOps-Copilot

Review copilot-api responses to ensure every answer cites metrics/logs evidence. Block any response that hallucinates or leaks system prompt.

2026-05-010

metrics-emulator-tdd

thc1006/OrbitOps-Copilot

Implement ntn-metrics-emulator features TDD-style — red commit (failing test) → green commit (impl) → refactor. Strictly per SPEC-002 + metrics.schema.json.

2026-05-010

name	runspace-pitch
description	Build / refine RunSpace 10-page deck, 90-second video, 3-minute English-subtitled video. Anonymity-enforced.

Overview

Triggers

Editing docs/06_runspace_pitch_outline.md, docs/07_demo_script_90s.md, docs/08_demo_script_3min.md.
Producing a slide deck (PDF) or video for submission.
Slash command /research produces a new fact that should be reflected on a slide.

Inputs

docs/01_product_strategy.md (UC1/UC2 + demo story).
docs/00_research_2026_04.md (sources + standards anchoring).
docs/03_breakthrough_directions.md (roadmap).
docs/10_links.md (allowlisted footnote URLs).

Step-by-step workflow

Open the matching outline / script doc.
Pull narrative beats from docs/01_product_strategy.md.
Anchor every external claim to a footnote citing a docs/10_links.md URL.
Cross-check against docs/00_research_2026_04.md confidence column — mark Medium/Low claims with hedge language.
Draft slide / scene; render in dry-run.
Anonymity sweep: scripts/check-no-secrets.sh over generated material.
(For final video / PDF) exiftool -all= <file> to strip metadata.
Run claims-audit skill against the new pitch text — every claim must classify cleanly.

Output format

Updated outline / script .md.
Optional tmp/runspace-deck.pdf and tmp/runspace-90s.mp4 (gitignored, but submitted externally).

Verification checklist

No team / school / personal identifiers in text, slides, video, or rendered metadata.
Every external claim has a footnote URL from docs/10_links.md.
Two-frame self-score updated if scoring narrative shifts.
./verify.sh clean (anonymity gate).
Video metadata stripped via exiftool -all=.
OS toolbar / browser tabs not visible in any screenshot.
claims-audit skill produced zero OVER-CLAIM rows.

Common failure modes

Slide footer auto-stamped with author identity from the slide tool (PowerPoint, Keynote).
Demo recording with terminal showing whoami / hostname in prompt.
A new claim added without a corresponding entry in docs/10_links.md.
"About us" slide accidentally drafted — must be removed entirely.

Forbidden actions

Naming team / school / individuals.
Quoting a metric / outcome that the implementation has not actually demonstrated (use claims-audit first).
Including any logo other than OSS project logos with attribution.
Using stock photography that has embedded metadata or watermarks.