跳到正文
cccskills
免费在 GitHub 公开

pb-social-pack

Turn a release or ship recap into a social pack: posts per platform where every claim traces to a facts file, clip cut-downs per aspect ratio cut with ffmpeg and checked with ffprobe, hand-off for publishing only after GO. Use for "social pack for the release", "posts and clips for this launch", "announce the release".

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md4.6 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Release social pack

Part of pb-content-studio (step f: social pack). What you get: posts per platform and clip cut-downs from your release recap, every claim traceable to a fact, handed off for publishing after your GO.

Inputs

  • recap — release notes, or a pb-ship / pb-release-aos run log
  • platforms — for example LinkedIn, X, Instagram, YouTube Shorts
  • launch_video (optional) — an mp4 from pb-launch-video; without it the pack is posts only and step 4 is skipped; pb-launch-video renders with hyperframes when installed, else with the remotion skill
  • formats (optional) — default 9:16 (1080x1920), 1:1 (1080x1080) and 16:9 (1920x1080)

Steps

  1. Preflight — test -d ~/.claude/skills/content-engine; absent → log "content-engine not installed: copywriting fallback" and continue. With launch_video: command -v ffmpeg and command -v ffprobe; either missing → stop with "Missing tool: <name>. Install ffmpeg and rerun." and write nothing except the run log — probes answered
  2. Ask — recap path, platforms, slug, launch_video, formats → run log — all answered, recap file exists
  3. quick-recap — recap → social/<slug>/facts.md, one numbered fact per line with its source line in the recap; no fact without a source — file written
  4. content-engine (if present) or copywriting — facts.md → social/<slug>/posts.md per platform, each claim tagged [F<n>] — every claim has a tag that exists in facts.md — stops for approval
  5. ffmpeg — launch_video → social/<slug>/clips/<name>_<WxH>.mp4 per format and per requested cut, for example ffmpeg -ss <start> -t <seconds> -i <video> -vf "scale=<W>:<H>:force_original_aspect_ratio=increase,crop=<W>:<H>" -c:a copy <out>, then ffprobe -v error -select_streams v:0 -show_entries stream=width,height -of csv=p=0 <out> — ffprobe equals the asked WxH for every clip; on a mismatch stop with "clip <file> has WxH, expected ...", do not publish
  6. [GO] publish — platforms, post texts and the full clip list shown. The run stops here until the human types GO. No publish MCP exists, so the GO releases the hand-off: posts, clips and captions in social/<slug>/publish.md; upload is manual unless the human named a tool in step 2 — hand-off file written

Run log: production_artifacts/pb-social-pack-<date>.md in the start directory, never committed

Optional silence-cut step: auto-editor

  • auto-editor (https://github.com/WyattBlue/auto-editor) cuts silent sections from a clip automatically — useful before scale/crop when the launch video is a talking-head recording.
  • Preflight probe: auto-editor --version. If the binary is missing or errors, log "auto-editor not installed: silence-cut skipped" and continue — the step never fails the run.
  • Usage when present: auto-editor <launch_video> -o <out.mp4> before the ffmpeg cut-downs in step 5; the cut-downs then take the silence-cut file as input.
  • Installing auto-editor (pip install auto-editor) is a manual opt-in step for the user — never pip-install it as part of this playbook.

Rules

  • Anything other than the literal GO (case-insensitive) is not a GO; a GO covers only that one step, one time.
  • A failed check stops the run: write the failure into the run log and report. No silent retries.
  • GO is by contract in every harness. The go-gate hook is only a backstop on Claude Code, OpenCode and agy (Codex: unverified; Cursor, Kimi: none). A missing hook is never permission to proceed.
  • Write one run-log line per step as it completes (N. done|skipped|failed — artifact — check result) and WAITING FOR GO: <step> at each gate.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Use when a hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Reference for how BDB structures autonomous software engineering work — the seven-node dispatcher graph (Architect, TechLead, UI/UX, Engineering, Media/EventTech, Reviewer, Shipping) that /startcycle-graph actually runs. Use when you need the high-level lifecycle framing without inventing your own process.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

agenttrail: live map of a multi-agent build in the browser: which plan component is being worked on, by which agent or harness, what is done and what is stuck. Use when a multi-agent pipeline starts (/startcycle, /startcycle-graph, /teamwork-preview) or after a plan-canvas approve, when the user asks to see what the agents are doing ("live map", "build board", "who is running now"), or by yourself whenever a multi-agent run is under way — no user prompt needed.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

查看 hybridlabor-api 的全部技能

报告此技能的问题