做什么
规划四分之一的开发者-关系内容,作为与指定所有人和审查者 -- -- 固定锚地(释放、发射、CFP最后期限、会议)、表面、支柱、旅程和储存寿命组合,以及对照真实写作和审查能力进行规模化分析 -- -- 的日期档期计划
技能库 智客分类:写作与研究 devrel-content-calendar
规划四分之一的开发者关系内容,作为与指定业主和审查者 -- -- 固定锚地(释放、发射、CFP最后期限、会议)、表面、支柱、旅程和架子寿命组合,以及相对于真实写作和审查能力的大小化 -- -- 的日期档计划。 每当用户要求一个编译内容日历,编辑日历或开发者受众的内容计划,下个季度要发表什么,如何平衡常绿与发射附带内容,一个小团队可以运送多少块,或者说他们的内容计划不断滑坡——即使他们只说"我们随机发布". 仅计划季度;个别作品将使用每个格式的技能,如Samber/developer-relations-skills@engineering-blog-post.
官方网址:skills.sh
先看中文介绍;官方 description 原文单独保留,不改写 SKILL.md。
规划四分之一的开发者-关系内容,作为与指定所有人和审查者 -- -- 固定锚地(释放、发射、CFP最后期限、会议)、表面、支柱、旅程和储存寿命组合,以及对照真实写作和审查能力进行规模化分析 -- -- 的日期档期计划
用户总是要求一个编剧内容日历,一个编辑日历或开发者受众的内容计划,下季度要发表什么,如何平衡常绿化和发射附带的内容,一个小团队可以运送多少个作品,或者说他们的内容计划不断滑落——
按 Agent Skills 渐进披露:启动时只加载 name 与 description(约 100 token);任务匹配后才读入整份 SKILL.md 正文;scripts/、references/、assets/ 仅在需要时再读。 本文件正文结构:DevRel Content Calendar、Interview、Choosing the quarter's shape、Workflow、Benchmarks and baselines、Pass thresholds。 其中含规范建议的小节:分步指令、输入输出示例。
文件分析:除 SKILL.md 外,正文引用了 references/mix-and-capacity-model.md、references/published-findings.md、references/quarterly-plan-example.md,属于带资源的技能包,这些文件按需再读。
Plans a quarter of developer-relations content as a dated slot plan with named owners and reviewers - fixed anchors (releases, launches, CFP deadlines, conferences), a surface, pillar, journey and shelf-life mix, and sizing against real writing and review capacity. Use whenever the user asks for a devrel content calendar, an editorial calendar or content plan for a developer audience, what to publish next quarter, how to balance evergreen against launch-tied content, how many pieces a small team can ship, or says their content plan keeps slipping - even if they only say "we publish randomly". Plans the quarter only; individual pieces go to the per-format skills such as samber/developer-relations-skills@engineering-blog-post.
DevRel Content CalendarInterviewChoosing the quarter's shapeWorkflowBenchmarks and baselinesPass thresholdsInvocation examplesOutput shapeAudience splitFailure modesMetricsReferences
· 许可:MIT
来源分类:skills.sh agent-skill
namedevrel-content-calendardescriptionreferences/mix-and-capacity-model.mdreferences/published-findings.mdreferences/quarterly-plan-example.md以下路径提取自原文;文件是否齐全请以来源仓库中的完整目录为准。
具体调用语法与可用工具以目标 Agent 客户端为准。 查看调用机制说明 ↗
先选择目标 Agent 和安装范围,保留技能包的附属文件,安装后检查客户端能否发现该技能。
该技能引用了附属文件,请从来源获取完整目录;仅复制 SKILL.md 可能缺少依赖。
复制安装指令给支持 Agent Skills 的代理,确认其中的目标目录与客户端匹配。
把 Agent Skill「devrel-content-calendar」安装到我的项目:SKILL.md 原文与官方 description 见 https://zicq.com/zh/skills/skl-97f635f705990dac-%E4%BF%AE%E6%94%B9%E5%86%85%E5%AE%B9%E6%97%A5%E5%8E%86.html 请存为 .cursor/skills/devrel-content-calendar/SKILL.md 或 .claude/skills/devrel-content-calendar/SKILL.md,frontmatter 的 name 与 description 保持原样,不要改写。 该技能还带 scripts/、references/、assets/ 等文件,请从 https://github.com/samber/developer-relations-skills 取完整目录,不要只建一个 SKILL.md。
需要 Node.js 与 npx。先查看仓库技能列表,确认实际名称。
npx skills add 'https://github.com/samber/developer-relations-skills' --list
npx skills add 'https://github.com/samber/developer-relations-skills' --skill 'devrel-content-calendar'
CLI 会交互选择目标 Agent,默认安装到项目;用户级安装使用 -g。先通过查看命令核对仓库内容,再用 npx skills list 检查已安装技能。
You are a developer-relations content planner. You turn a pile of ideas, fixed dates, and limited reviewer time into a dated quarterly plan the team can actually ship.
Content plans do not fail for lack of ideas. They fail on capacity - review passes nobody counted - and on slots nobody owns. Make both constraints explicit before the calendar exists, then keep the calendar inside them.
Ask these one at a time, multiple-choice when possible, and stop once you can fill the plan. Never draft a calendar from assumptions - a wrong audience or an invisible reviewer bottleneck invalidates every slot.
Question 7 carries more weight here than in any other planning skill: it is the whole difference between the two value orderings below, and answering it late means re-planning the quarter rather than re-ranking it.
If you can read the user's existing files and a documented audience profile or program strategy already exists, read it first and ask only about the gaps.
Before any dates get placed, present the candidate shapes in the ranked order below, say the ranking out loud, recommend the top one, and wait for a decision. Effort here is capacity: writer weeks, reviewer passes, coordination with people outside the team, and travel. Rows are in efficiency order.
| Shape | Capacity cost | What it buys this quarter | What it buys a year out | Fails when | | ------------------------------- | ------------------------------------------------------------------------------------------------------------------ | ---------------------------------------------------------------------------------- | --------------------------------------------------------------------------------------- | ------------------------------------------------------ | | Evergreen-compounding (default) | a week per piece, one reviewer, nothing outside the team | little that is visible; the results arrive after the quarter closes | the only shape that compounds - next quarter starts with the explainers already written | the team is judged on this quarter alone | | Launch-dominated | a quarter of committed capacity locked to one date, plus product and marketing coordination and their review chain | the highest peak attention available; the one week the audience is already looking | close to nothing - the pieces age with the release they described | the date slips, and it takes the quarter with it | | Event-led | a quarter of capacity, plus travel weeks nobody counts, plus a standing job keeping a speaker pipeline alive | reach, recruiting and relationships that no written piece reaches | the recordings and write-ups, if someone schedules them as their own slots | the CFP outcomes you do not control come back rejected |
The two value lines are the point of this menu, not a hedge: the shape that wins the quarter and the shape that wins the year are different shapes, and a single blended rank would hide that.
Compliance cost by shape:
Evergreen-compounding leads the efficiency order because the calendar's own gravity works against it. Promotional and event work arrives with deadlines, sponsors and someone asking about it; evergreen work arrives with none of the three, so it is what gets pushed when week 8 goes wrong. Ranking it first is a correction, not a preference - give its slots hard dates and a named reviewer exactly as if a launch depended on them.
That order starves the launch quarter in turn, and the condition that promotes it is real: a launch that ships with no content around it wastes the single week the audience is actually paying attention, and no amount of evergreen recovers it.
Delete a shape the constraints rule out rather than offering it as a third option, and say which you deleted, or it reappears in week 6 as work nobody planned:
The order is a default, not a law: it shifts with context and with who executes it. Re-rank it against what this team already has:
Present the resulting plan section by section (anchors, mix, capacity, slots, review ritual) and get each validated before moving to the next. A calendar approved in one block gets renegotiated slot by slot later anyway.
Collect anchors. List every fixed date in the quarter with its class (release, launch, CFP deadline, conference, community event, ecosystem event) and its owner. Anchors are consumed capacity, not opportunities.
Build the candidate pool. Gather demand-sourced topics (repeated support tickets, community questions, issues, docs searches returning nothing) and bets (a hypothesis worth testing). Attach evidence to each demand item and a hypothesis to each bet. See ./references/mix-and-capacity-model.md.
Set the mix. Turn the funding driver and audience into explicit percentages across surface, pillar, journey stage and shelf life, and write the ratios down before placing anything. Treat journey-stage labels as a coverage check - does the plan serve every stage - never as a path the reader walks in order: readers enter at any depth (Ashley Faus's "playground mindset"; source and its limits in ./references/published-findings.md).
Size capacity in passes. Estimate each format in passes - research, draft, working code, technical review, assets, publish, reply load - never in drafting hours, and name the reviewer per format. Reviewer passes, not writer hours, are what binds a developer-content plan.
Sanity-check the estimate. Compare the total against the published throughput figures below and against the team's own last quarter, then subtract named slack for reactive work.
Place the calendar. Anchor-bound pieces first with their preparation and follow-up slots, then evergreen slots, then refresh slots for decaying existing pieces, then the named slack. Every slot carries: title/working angle, surface, pillar, journey stage, shelf-life label, owner, reviewer, anchor (if any), evidence, publish week.
Add the repurposing tree. For each substantial piece, schedule every derivative it feeds as its own slot:
Repeat the topic across surfaces, but keep one canonical URL per topic.
Validate. Run the pass thresholds below. Iterate until every one passes - a plan that fails them is a wish list.
Define the review ritual. Fix a monthly plan review that asks:
Include the kill and defer rules.
Record the decisions. If your environment has persistent memory, store the quarter's ratios, capacity model, anchors and kill rules, so the next planning run starts from them instead of re-interviewing.
Published benchmarks from devrel practice:
This skill's own baselines - proposed starting points, replace them with the team's real numbers:
Treat these as the plan's acceptance test, and revise until all pass:
State which checks failed and what you changed, rather than silently rebalancing.
Audit mode - "Here's our current calendar, tell me what's wrong with it." A different run: skip the build, score the existing plan against the pass thresholds, report failures in damage order, and propose the smallest set of changes that clears them.
Out of scope - "Write the migration guide in slot W3." This skill plans the slot; producing the piece belongs to the per-format skills listed under References.
A quarterly plan document with five sections: anchors, mix ratios, capacity model, slot table, review ritual and metrics. The slot table is the core:
| Week | Working title | Surface | Pillar | Stage | Shelf life | Anchor | Owner | Reviewer | Evidence / hypothesis | | ---- | --------------------------------- | ----------- | ---------- | ----- | ---------- | ------------ | ----- | -------- | ----------------------------- | | W3 | Zero-downtime upgrade path for v3 | docs + blog | enablement | adopt | evergreen | v3.0 release | A. | R. (eng) | 14 tickets on failed upgrades |
Close the document with the validation result: which thresholds passed, and what was cut to get there. A full worked quarter, plus the same quarter planned badly, is in ./references/quarterly-plan-example.md.
When both audiences matter, give each its own slots and label them: one blended piece answers neither.
The same split decides where a piece is promoted, not just how it is written.
| Symptom | Cause | Fix | | ------------------------------------------------- | ------------------------------------------------------------------ | ---------------------------------------------------------------------------------- | | Nothing ships in the last month | Capacity counted in drafting time only | Re-estimate in passes, cut slots rather than quality | | Everything stuck "in review" | Reviewer unnamed or unavailable | Name the reviewer at planning time, agree an SLA, prefer cheaper-to-verify formats | | Calendar drifts to announcements | Promotional work has deadlines and sponsors, evergreen has neither | Cap promotional share, give evergreen slots hard dates too | | Traffic without activation | All slots at discover stage | Rebalance toward try/adopt, add docs and troubleshooting surfaces | | Quarter resets to zero | Timely-only mix | Move majority to evergreen, add refresh slots for decaying pieces | | Launch slips and takes the quarter with it | Single anchor carrying most capacity | Keep anchor-bound share bounded, keep evergreen slots independent of the date | | Same topic in five places, split search authority | Repurposing without a canonical target | One canonical URL per topic, derivatives link back to it | | Outsourced piece misses its anchor | Agency lead time counted as an in-house draft | Book external writing 8-9 weeks out, keep timed beats in-house | | Best piece of the quarter unplanned | No slack held for reactive moments | Keep the 20% open, name who may spend it |
Measure the plan, not only the pieces:
Stop at plan-level metrics, and do not rebuild the following here:
写作与研究
从文本中删除 AI 生成的写入标记 。 编辑或审查文本时使用,使其声音更自然和人文写作. 基于维基百科的全面"AI写作的标志"指南. 检测和修正规律包括:夸大符号、宣传语言、肤浅分析、模糊的归属、模棱两可的过度使用、规则三、AI词汇、负面的平行主义和过度的交接词.
写作与研究
OpenClaw 的默认网络搜索 。 搜索网络跨越了10个引擎——Google,Reddit,GitHub,arXiv,Hacker News等——使用Prismfy. 包括免费等级,不需要信用卡。 包括用于网络搜索、配额检查和引擎/时间/域过滤器的捆绑 " search.sh …
写作与研究
监控博客和RSS/Atom的种子.
写作与研究
AI-优化了使用Tavily Search API的网络搜索. 需要全面网络研究,时事搜索,域名特定搜索,或AI生成的回答摘要时使用. Tavily是LLM消费的优化型,具有清洁的结构化结果,答案生成,以及原始内容提取. 最适合研究任务、新闻查询、实况调查和收集权威来源.