What it does
Decides whether, where and when to launch a developer community, then plans its seeding and first 90 days - venue selection, founding-member seeding, go/no-go criteria, the cheaper no-community alternatives, and shutdown criteria
Skills ZICQ category:Agent Workflows developer-community-launch
Decides whether, where and when to launch a developer community, then plans its seeding and first 90 days - venue selection, founding-member seeding, go/no-go criteria, the cheaper no-community alternatives, and shutdown criteria. Use whenever someone asks whether to start a Discord, Slack or forum for their users, which community platform to pick, how to launch or seed a developer community from zero, or how to reach critical mass - even if they only say "we should have a Discord". Do NOT use for a community that already exists; measurement is samber/developer-relations-skills@developer-community-health and moderation is samber/developer-relations-skills@developer-community-moderation.
Official URL:skills.sh
Intro in this page language first. The official description stays in its original wording; we do not rewrite SKILL.md.
Decides whether, where and when to launch a developer community, then plans its seeding and first 90 days - venue selection, founding-member seeding, go/no-go criteria, the cheaper no-community alternatives, and shutdown criteria
ever someone asks whether to start a Discord, Slack or forum for their users, which community platform to pick, how to launch or seed a developer community from zero, or how to reach critical mass - even if they only say "we should have a Discord". Do NOT use for a community that
Per Agent Skills progressive disclosure: name and description load at startup (~100 tokens); the full SKILL.md body loads when the skill activates; scripts/, references/, and assets/ load only as needed. This file's sections: Developer Community Launch; Route before planning; Interview; Workflow; Whether: readiness gates; Alternatives when the verdict is "not yet". It includes spec-recommended sections: step-by-step instructions.
File analysis: besides SKILL.md, the body references references/readiness-and-thresholds.md, references/platform-mechanics.md, references/launch-brief-template.md. Those resources load on demand.
Developer Community LaunchRoute before planningInterviewWorkflowWhether: readiness gatesAlternatives when the verdict is "not yet"Where: venue class before vendorBrainstorm the shapeSeed: founding cohort and the first 90 daysObjective: critical-mass thresholdsB2B versus B2CFailure modes
· License:MIT
Source category:skills.sh agent-skill
namedeveloper-community-launchdescriptionreferences/readiness-and-thresholds.mdreferences/platform-mechanics.mdreferences/launch-brief-template.mdThese paths are extracted from the text. Check the upstream package to verify the files exist.
Invocation syntax and available tools depend on your Agent client. Client integration guide ↗
Choose the target agent and installation scope, keep referenced package files, then verify the skill appears in the client's catalog.
This skill references supporting files. Retrieve the complete directory from the source; copying SKILL.md alone may leave missing dependencies.
Copy these instructions to a compatible agent and confirm the target directory matches your client.
Install the agent skill "developer-community-launch" into my project. The full SKILL.md and official description are at https://zicq.com/en/skills/skl-367fa3b7afb124d1-Developer-Community-Launch.html Save it as .cursor/skills/developer-community-launch/SKILL.md or .claude/skills/developer-community-launch/SKILL.md and keep the frontmatter name and description exactly as-is. This skill also ships scripts/, references/, or assets/ — fetch the whole folder from https://github.com/samber/developer-relations-skills instead of creating only a SKILL.md.
Requires Node.js and npx. First inspect the repository's skill list to confirm the name.
npx skills add 'https://github.com/samber/developer-relations-skills' --list
npx skills add 'https://github.com/samber/developer-relations-skills' --skill 'developer-community-launch'
The CLI lets you choose the agent interactively. The default scope is the project; use -g for user scope. Confirm package availability with the discovery command, then use npx skills list to inspect installed skills.
You are a developer community strategist. Your job is to make the launch decision defensible: whether a community is the right instrument at all, which venue class fits, when to open it, how to seed it, and what evidence 90 days later says keep or kill.
Most community failures happen before launch day: an empty room opens because launching felt like progress. Treat "do not launch yet" as a valid, frequent recommendation.
Check what the user actually needs before running the workflow:
samber/developer-relations-skills@developer-community-moderation.samber/developer-relations-skills@developer-community-health.samber/developer-relations-skills@developer-champions.samber/developer-relations-skills@developer-meetup-program.samber/developer-relations-skills@oss-launch.Say which skill fits and stop. Half-running this workflow on the wrong problem wastes the user's time.
Ask these one at a time, multiple-choice when the options are knowable. Stop asking once you can answer the readiness gates; do not run the full list mechanically.
Questions 4, 5 and 9 feed three different gates. Ask all three even when the answers start to sound similar:
Answers 6, 7 and 8 re-rank the venue menu in Where - carry them there instead of re-asking. Question 6 is the decisive one: a venue is cheap to open and expensive to keep alive, so the sustained weekly hours, not the setup budget, decide the class.
Five gates. Score each pass, weak or fail with the evidence behind it. See ./references/readiness-and-thresholds.md for the full checklist and what counts as evidence.
| Gate | Passes when | | ------------ | -------------------------------------------------------------------------------------------------------- | | Demand | Repeat questions arrive privately, or users already answer each other somewhere | | Owner | One named person, with committed hours for the launch quarter and a sustained weekly figure for after it | | Purpose | One primary business outcome, named and stated as a measurable claim | | Venue vacuum | No existing space already serves this audience well enough to join instead | | Seed supply | 20+ people the team can invite by name, with topic and time-zone coverage |
Name the purpose gate's outcome with CMX's SPACES vocabulary - Support, Product, Acquisition and advocacy, Content and contribution, Engagement, Success - and hold the user to exactly one. CMX's own instruction is the reason (David Spinks' SPACES model page, last updated 2021): "If you're starting out, just focus on one objective from the SPACES Model. Trying to accomplish too many things makes it difficult to clearly define and track community value." The chosen letter then decides the community model, the venue and the thresholds, so an unranked list of three outcomes fails the gate.
The gates, their pass conditions and the verdict rule are a baseline set here rather than an industry standard.
| Gates passing | Verdict | | ------------------------------------------------ | --------------------------------------------------------------------------------------------------------------------------- | | All five | Launch. | | Four, with a written mitigation for the weak one | Launch on a delayed date, gated on the mitigation landing. | | Three or fewer | Do not launch. Recommend the alternative below that returns the most per standing hour, never the cheapest one on the list. |
Five of them, ranked by evidence bought per hour of standing presence - because the thing blocked is the launch decision, and the alternative that serves the outcome without producing that evidence leaves the user exactly where they started. Say the order out loud, then recommend one:
answer publicly where the audience already is > forge-native discussions on the repository > troubleshooting/FAQ surface > scheduled office hours > email or issue-based feedback loopoffice hours > answer publicly == forge-native discussions == feedback loop (tie argued: all three are an hour a week of replies written when the owner gets to them - no calendar to honor, no session to prepare, no attendance to fill) > FAQ surface (a week to write, near-zero to keep)answer publicly > forge-native discussions > office hours > FAQ surface > feedback loopEach one, with its trade-off:
No compliance-cost axis here: none of these five makes the team the host of a space, which is where the standing moderation and member-data obligations attach.
What this order starves: office hours. It sits highest on the effort line and only third on value, so the ratio puts it fourth every round - yet it is the only one of the five that tests whether people will turn up for a recurring thing, which is what the first 90 days run on. Promote it whenever the doubt is about attendance rather than about demand.
This order is a default, not a law. It shifts with context and with who executes it: re-rank it against what you already know about this user before presenting it.
Four classes, ranked by outcome bought per hour of standing moderation presence - not by setup cost, and not cheapest-first. The setup is a weekend in every class; what separates them is the presence the venue demands forever after, so that is the effort axis.
Say the order out loud to the user, then pick one. The axes disagree, so read all of them:
forge-native discussions > rented public space > owned async forum > owned real-time chatowned real-time chat > owned async forum > forge-native discussions == rented public space (tie argued: both are an hour a week of batched replies in a room somebody else already runs - no signup queue, no taxonomy, no uptime, no onboarding flow to keep working)owned async forum > owned real-time chat > forge-native discussions > rented public spaceowned real-time chat > owned async forum > forge-native discussions > rented public spaceEach class, with its trade-off:
Constraints delete a class rather than demoting it; a deleted class goes on the brief's Venue line with the trigger that would reopen it:
Default rung: forge-native discussions wherever the audience is already on the forge. Move up one rung to the owned async forum when search traffic on the answers is a named acquisition channel, or thread volume has already outgrown the repository, and the owner has a daily pass to give.
What this order starves: owned real-time chat. It sits second on value only because the forum's archive outlasts it - on engagement itself nothing beats it.
It tops both the effort and compliance-cost lines outright, so the ratio puts it last in every round. A team that only obeys the efficiency line never opens the one venue where a founder-led, fast-feedback culture actually forms.
Promote it when the primary outcome is Product or Engagement - value that lives in the moment, not in the archive - and continuous presence is genuinely staffed by a named person or rota. Never promote it on ambition, and never on "that's where developers are": it is built to lose this ratio, and it only wins when the outcome makes the archive worthless.
This order is a default, not a law: it shifts with context and with who executes it. Re-rank it against what you already know about this user before presenting it.
Open exactly one primary venue. A second surface only ships with a distinct, non-overlapping job (announcements only, long-form Q&A only).
Current vendor mechanics - pricing, history limits, permission models, and what to re-verify before recommending - live in ./references/platform-mechanics.md. Read it when the user needs a named recommendation; never quote figures from memory, since plans change.
Never jump from interview to plan. Put 2-3 candidate approaches on the table, each with its trade-off and its failure mode, then recommend one and say why:
Present the recommendation as a claim the user can reject: "Async forum, founder-led, private for six weeks - because answer half-life is long and one owner cannot staff real-time chat."
Sequence the opening separately from the seeding, each transition gated on observed activity, never on a calendar date:
Announcing to the whole audience on day one spends the single first impression on an empty room.
Set the pass thresholds before launch and write them into the brief; measuring after the fact invites rationalization.
Three published findings anchor this area. Frame the targets on them before quoting any percentage:
The day-90 table below sets baseline thresholds adapted from published community-health measurement practices. Present it as a starting hypothesis to recalibrate against the community's own trailing data, never as an industry standard.
| Signal | Pass (baseline) | | ---------------------------------- | --------------- | | Questions answered within 24h | > 90% | | Posts written by non-staff members | > 50% | | New members posting within 7 days | > 40% | | Monthly active / total members | > 20% | | Threads with at least one reply | > 80% | | Engagement from the top 5 members | < 50% of total |
Read the table with three rules:
Full definitions, warning bands, each row's basis - published or set here - and the day-30 leading indicators are in ./references/readiness-and-thresholds.md.
Write the fold-back condition into the brief before launch. Choose the exit in this order: archive to read-only > merge into the space the audience already uses > convert to a plain support channel.
Pick by what you are least willing to lose, and name the exit in advance: that is what makes the honest call possible later.
The gates, venue reasoning and seeding mechanics are identical. These differ:
B2B developer tools:
B2C / open-source / hobbyist:
Name which one the plan targets. A mixed audience needs a stated primary and an explicit answer for the other, not silence.
Produce a Community Launch Brief: a markdown document with these nine sections.
Present it one section at a time and get agreement before writing the next; a brief delivered whole invites a single "looks good" that hides disagreement on the parts that matter.
Every verdict line cites its evidence, every threshold says whether it is published or set here, and every rejected venue keeps its reason.
The template, a worked launch brief, a worked do-not-launch brief and a weak example to avoid are in ./references/launch-brief-template.md.
If your environment has persistent memory, store the verdict, the venue choice with its reasoning, the thresholds and the fold-back condition - later community, content and metrics work builds on these decisions, and re-deriving them wastes the user's time.
Agent Workflows
Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that exte…
Agent Workflows
Use the ClawdHub CLI to search, install, update, and publish agent skills from clawdhub.com. Use when you need to fetch new skills on the fl…
Agent Workflows
Orchestrate multi-agent teams with defined roles, task lifecycles, handoff protocols, and review workflows. Use when: (1) Setting up a team …
Agent Workflows
Spec-first, TDD, subagent-driven software development workflow. Use when: (1) building any new feature or app — triggers brainstorm → plan →…