trainingsites/campus-ai-os/brief-builder
Claude Code and agent skill by trainingsites/campus-ai-os
- community
Checking that the SKILL.md file is still in the repository.
Builds the spec-conformant Content Brief (brief.md + brief.json) for the Strategy and Research Team (Atlas) from top-scored topics. This brief is the FRONT SOCKET every execution team reads. Internal skill, invoked by content-scanner/atlas-daily.
Install
npx vskill@latest install trainingsites/campus-ai-os/brief-builderSKILL.md
brief-builder — emit the front socket
Produces a Content Brief (.md + .json) that conforms to Team Interop Spec v1. Every execution team reads this — it must be complete and portable.
CRITICAL DIRECTIVES
- Emit `brief.json` (machine source of truth) AND `brief.md` — never only the markdown. Write both to the briefs location:
./teams-shared/briefs/{slug}/standalone,outputs/{YYYY-MM}/briefs/{slug}/under a host campus. Never inside a run folder. (This is the #1 interop bug — the whole point of Atlas is thebrief.jsonother teams read.) - Do not invent a format. Use
references/brief-schema.mdexactly. It is the shared brief family every execution team reads — never fork it. run_id+schema_versionare REQUIRED on every brief. run_id is the §4 format, worked example20260530-ai-operating-system-fo-7f3a.- Tool-agnostic; zero hardcoded paths. Voice/audience resolve via
CONTEXT_CANON(.campus-os/context-files.md, else the plugin'stemplates/context-files.md) (canon), never hard-codedshared/*.
Step 1 — Inputs
Take the top-scored topic(s) from scoring-engine plus the scanner's signal evidence and any mandate.json. Get the run_id from atlas-orchestrator (worked example 20260530-ai-operating-system-fo-7f3a; inherit if present, never re-mint).
Step 2 — Choose output_type
Map the topic + business_model defaults to one of the 16 output types (youtube_script, social_bundle, blog_post, webinar, workshop, presentation, course_module, email_campaign, brief_only, …). If the user already named a target team/format, honor it.
Step 3 — Build the object
Fill the required spine + scoring block (from scoring-engine; null in pure standalone) + production payload. For writer/teaching lanes, ensure key_messages, quotable_moments, and content_structure are present (enrich from sources if thin). tags = portable plain strings only.
Step 4 — Write both files
Generate a slug (lowercase, hyphenated, ≤60; append -2/-3 on collision). Write to the handoff location:
- Connected →
outputs/{YYYY-MM}/briefs/{slug}/brief.{md,json}(mirror to./teams-shared/briefs/{slug}/if peer teams expect it). - Standalone/peer →
./teams-shared/briefs/{slug}/brief.{md,json}.
Set status: idea, today's dates, schema_version: "1.0", and the inherited run_id.
Step 5 — Route
Hand to output-router: deliver the brief and tell the user which execution team to run next (or generate the standalone brief if no downstream team is installed).
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/brief-builder/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/trainingsites/campus-ai-os/brief-builder)