samber/dev-event-organizer-skills / skills/event-speaker-sourcing
event-speaker-sourcing
VerifiedFind and qualify speakers for a technical event outside the open call for papers - the proactive channel running alongside or instead of a CFP. Covers program gaps worth filling by invitation, sourcing channels (referral asks, expert and speaker directories, prior-talk review, community scouting), international sourcing by region and language, evidence-based Hot/Warm/Cold/Skip qualification, and a ranked shortlist carrying the "why" behind each name. Use whenever the user mentions finding speakers, a keynote or invited-speaker shortlist, sourcing international or under-represented speakers, or a programme a CFP did not fill - even if they never say "sourcing". Do NOT use to write the invitation - use samber/dev-event-organizer-skills@event-speaker-cold-outreach instead.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 29.7 KB |
| evals/evals.json | config | 29.6 KB |
| references/shortlist-and-qualification-template.md | markdown | 8.1 KB |
| references/sourcing-channel-evidence.md | markdown | 6.9 KB |
evals/evals.json
View on GitHub{
"skill_name": "event-speaker-sourcing",
"evals": [
{
"id": 1,
"prompt": "NordStack Days is our single-track community conference in Oslo, 10 speaking slots, and our CFP closes in 12 days. We have 4 proposals worth accepting. I pulled 6 names from my old work Slack - people I've worked with who'd say yes - and I also made a spreadsheet of the platform and infra accounts with the biggest followings on tech Twitter, sorted by follower count. Can you turn that into a ranked invite list so I can start emailing tonight?",
"expected_output": "Identifies this as the supplement-to-a-live-CFP trigger, refuses to rank on follower count, requires slot briefs before names, adds a second channel that reaches outside the organizer's network, and stops at a validated shortlist rather than an email.",
"files": [],
"expectations": [
"Names the trigger explicitly as a supplement to a live call for papers, not the replacement case where the call was deleted",
"States that thin early CFP volume is ordinary, expected practice rather than treating it as a crisis",
"Refuses to rank or score candidates by follower count or audience size",
"States that audience-size metrics are excluded from the qualification rubric entirely, rather than kept as a lower-weighted factor or a tiebreak between otherwise equal candidates",
"Requires a written slot brief for each gap before any candidate name is attached to it",
"Specifies the slot brief contents as at least theme, session format, audience level, and what success looks like for that slot",
"Requires at least two sourcing channels, at least one of which reaches beyond the organizer's own network",
"Identifies the six Slack names as the referral-asks channel and states its ceiling is that it returns the organizer's own network's network",
"Declines to write the invitation message itself and routes that work to samber/dev-event-organizer-skills@event-speaker-cold-outreach",
"States the shortlist is presented slot by slot for the organizer's validation before any invitation is sent",
"Requires delivery evidence - a recorded talk, a transcript, or long-form writing - for every name that reaches the shortlist",
"States that the size of the gap is re-checked after the call closes, since the final proposal count is not yet known"
]
},
{
"id": 2,
"prompt": "Kettlebridge Infra Day, first edition, 12 speaking slots. We killed our call for papers because we ran out of runway - the event is 8 weeks from today and the whole programme now has to be invited. Budget is flights and hotel covered, free ticket, and zero speaker fee. That's confirmed and it's not moving. A colleague says we should just pay a speaker agency to fill 4 of the slots quickly - worth it? My own plan was to spend the next three weeks going through conference videos on YouTube to pick the best 12 people. What order should I do this in?",
"expected_output": "Names the replacement trigger, deletes paid bureau booking outright because no fee is confirmed, re-ranks channels toward fast vouched ones for the deadline, and works back a confirmation date that respects the 4-6 week travel floor.",
"files": [],
"expectations": [
"Names the trigger as the replacement case, where a deleted call for papers means the entire lineup is sourced by invitation",
"Deletes paid speaker-bureau or agency booking from the channel menu outright rather than ranking it low or keeping it as a fallback",
"Gives the reason for that deletion as the absence of a confirmed fee, since a bureau's mechanism is a negotiated fee",
"Warns that an option parked at the bottom of a menu silently returns as scope partway through sourcing",
"Re-ranks the channel menu against the hard deadline, promoting vouched, fast-answering channels over thorough ones",
"Places referral asks at or near the top of the channel order for this timeline",
"Gives the reason that a referral arrives vouched and therefore skips a qualification step an unvouched name still has to pay for",
"Rejects three weeks of prior-talk review as the primary plan for 12 slots, because it costs real hours per candidate",
"Keeps prior-talk review in the plan for at least the highest-risk slot rather than dropping the channel entirely",
"States the 4-6 week speaker travel lead time as a floor and minimum, not a target",
"Works back from the 8-week event date to a speaker confirmation deadline roughly 2 to 4 weeks from today so the travel floor still fits",
"States the lead-time gate applies when a candidate enters the shortlist, not when outreach happens",
"Still requires at least two channels rather than collapsing the plan to referrals alone"
]
},
{
"id": 3,
"prompt": "Meridian Container Summit is in Lisbon, 14 weeks out, 8 slots. Three of them we want filled by speakers from outside Europe - specifically Latin America and East or Southeast Asia - and everyone presents in English. We have flights confirmed in the budget for two of those three; the third is probably fine, we'll see where we land. Speakers have to be confirmed 5 weeks from now for the programme announcement. My plan is just to search harder on LinkedIn and Google until I find people. How do I actually do this?",
"expected_output": "Replaces 'search harder' with directory filtering on region and language together, gates every travelling candidate on visa lead time against the 5-week confirmation date, and forces the third travel budget to be stated as unconfirmed in the handoff.",
"files": [],
"expectations": [
"States the mechanism for international sourcing is filtering on published directory fields, not searching harder or longer on general-purpose search engines",
"Requires region and language filters to be applied together in the same query rather than separately",
"States that filtering on region alone returns candidates who cannot be briefed in a shared language",
"States that filtering on language alone returns candidates from the organizer's own country again",
"Names at least one live directory that publishes region, language, and topic or project filters usable together",
"Treats visa and entry requirements as a per-candidate check against the actual route rather than assuming one processing time",
"States that some entry routes need months of lead time and a formal invitation letter",
"Scores any candidate whose visa answer would arrive after the 5-week confirmation date as Skip for this edition",
"Flags that same candidate as a strong lead to carry into the next edition rather than discarding them",
"Records the third travel budget as unconfirmed rather than as covered, likely, or probable",
"States that an unconfirmed travel budget means that candidate is being asked to fund their own trip",
"Writes the unconfirmed budget status into the handoff so the invitation discloses it, rather than leaving it to surface during negotiation"
]
},
{
"id": 4,
"prompt": "Pelagic Data Conf, community-run, second edition, 6 invited slots. Two things I need numbers for so I can finish the budget spreadsheet. One: how many people do I approach to land 6? I've seen 2-3x quoted for outreach so I was going to send 15-18 invitations. Two: what do I put in the honorarium line per speaker? I assume a keynote is worth around $1,500 and a regular session less. Give me the figures.",
"expected_output": "Refuses both numbers, names the 2-3x multiplier as a borrowed sales figure, states community technical events have no standard honorarium rate, and redirects to the organizer's own prior edition plus their own legal and tax review.",
"files": [],
"expectations": [
"States the invitation-to-confirmation rate for invited speakers is unknown and no defensible general multiplier exists",
"Identifies the 2-3x figure as a lead-generation or sales number rather than a speaker-sourcing benchmark",
"Explains that the ask, the relationship, and the cost of a no all differ between sales prospecting and a speaker invitation",
"Labels the 2-3x figure as a shape for thinking if it is repeated at all, rather than applying it to produce a candidate count",
"Directs the organizer to size the long list from their own first edition's invitation outcomes",
"Instructs the organizer to record this edition's observed long-list-to-confirmation ratio as the baseline for the next one",
"Refuses to supply a market rate, a per-minute figure, or an industry-standard honorarium for an invited speaker",
"States that community technical events have no standard honorarium rate",
"Presents any published per-speaker figure it cites as calibration rather than a rate card, noting such figures disagree by more than 2x and sit in different currencies",
"Notes that published flat figures pay every accepted speaker the same amount regardless of format, which is a different shape from a negotiated invited-speaker fee",
"Warns that direct payment to speakers raises tax, visa, and conflict-of-interest or employment-agreement issues",
"Gates any fee policy on the organizer's own confirmed budget and their own legal and tax review"
]
},
{
"id": 5,
"prompt": "Halcyon Cloud runs our own user conference, 20 sessions, and we invite basically all of it. Two decisions stuck. First: our platinum sponsor, a monitoring vendor, has offered to provide a speaker for the 40-minute main-stage slot as part of their package, and honestly it would save me a lot of work. Second: for the cost governance session I can have a solutions architect from a FinOps tooling vendor who is a genuinely polished presenter, or a staff engineer at one of our customers who ran the whole migration but has never spoken at a conference in her life. Which way do I go on each?",
"expected_output": "Treats invited sourcing as the main intake at a vendor conference, refuses the sponsorship-to-stage conversion while still qualifying the sponsor's name on evidence, and screens the customer practitioner ahead of the vendor presenter without disqualifying the vendor outright.",
"files": [],
"expectations": [
"States that at a vendor-run user conference this sourcing pass is the main content intake rather than a backup to an open call",
"Weights the candidate pool toward customers, partners, and practitioners who use the vendor's technology",
"Refuses to let the platinum sponsorship package convert into the main-stage speaking slot",
"Requires the sponsor-supplied name to be qualified on the same evidence as any other candidate, or excluded from the shortlist",
"Treats an unresolved conflict with the programme's independence as a Skip disqualifier rather than a low score",
"Screens the customer staff engineer before the vendor solutions architect for the cost governance session",
"States a vendor-affiliated candidate fills the slot only when no practitioner is available, or appears as a second voice rather than the only one",
"Routes the vendor candidate to a workshop-or-niche shortlist or to a monitor-for-later list rather than deleting them",
"Requires delivery evidence for the customer engineer despite her having never spoken, accepting long-form writing or a transcript in place of a recorded talk",
"Declines to treat presentation polish alone as qualification for the vendor candidate",
"Notes the invitation at a vendor conference can usually carry real budget, unlike the community-run case"
]
},
{
"id": 6,
"prompt": "Tessellate Conf has one opening keynote slot. We're paying EUR 3,000 plus flights for it, that's signed off. I found three candidates through speaker directories, all three have decent recorded talks online, none of them I know personally and nobody I know has hosted them. Our programme committee already runs blind review on the CFP pool, so I figured we'd drop the keynote proposal into the same review queue to keep it fair. How do I pick between the three, and what do I hand to the person who writes the invitation?",
"expected_output": "Blocks the invited proposal from the blind-review pool and forces an explicit invited lane, promotes organizer reference for a paid keynote, and specifies the handoff packet contents.",
"files": [],
"expectations": [
"States an invited speaker's proposal cannot go through blind review with the open CFP pool, because their identity is the entire reason they were invited",
"Requires the invited-review lane to be decided now and written into the handoff packet rather than improvised later",
"Offers the two lanes as either confirmed at invitation and entering the programme directly, or reviewed on its own named track against the same criteria",
"Promotes organizer reference from a starved option to a required qualification rung for this slot",
"Gives the promotion reasons as the keynote or headline slot and the candidate being paid and flown at real cost",
"States that organizer reference is the only rung that predicts reliability rather than talent",
"Names the two failure modes it guards against: a great talk that never gets written, and a speaker who cancels late",
"Flags organizer reference as carrying the highest compliance cost of the qualification rungs, because it is a discussion about a named person with a third party",
"Constrains the reference conversation to what the referring organizer personally observed, and to statements the organizer would repeat to the candidate",
"Promotes prior-talk review ahead of directory search here, because a single slot concentrates the risk and the three names arrived unvouched",
"Ranks the three candidates within the keynote slot rather than across the whole programme",
"Lists the handoff packet contents, including the slot brief, the per-candidate why with dated evidence, the Hot/Warm/Cold/Skip score, availability and travel constraints, the honest budget offer, and the intended contact route"
]
},
{
"id": 7,
"prompt": "Bitfold is our community conference. The board set a goal that half the invited lineup should be first-time or under-represented speakers, and our co-chair Aneta owns that number. Our topic is narrow, WASM component model tooling, so there's maybe a few hundred people in the world doing it seriously. My plan: write a script to pull the top 200 contributors from the main repos, DM all of them in the project's Discord, and work out which ones are under-represented from their profile photos and names. I'll use speakerline.io as the speaker directory for the rest. Sound right?",
"expected_output": "Confirms community scouting is the right channel here while rejecting the scrape, the in-channel DMs, and the inference from photos and names, and corrects the speakerline.io assumption.",
"files": [],
"expectations": [
"Promotes community scouting for this programme and names the conditions met: a stated under-represented-speakers goal with a named accountable owner, and a topic too specific for any directory to cover",
"Rejects the bulk-pull script, stating that sourcing runs on public channels with no bulk scraping and no gated-data workarounds",
"Requires the community space's own code of conduct or rules to be checked before any names are taken from it",
"States that many community codes of conduct treat recruiting or solicitation inside the space as off-limits",
"Directs contact through each person's own published contact route rather than through the project's Discord",
"Refuses to infer under-represented status from profile photos or names",
"Replaces that inference with sourcing from a directory whose members opted in to being found for that reason",
"Names at least one live directory whose speakers self-register and that states organizer sourcing as its own purpose",
"Corrects the speakerline.io assumption, identifying it as a speaker-side CFP-transparency tool rather than an organizer-facing directory with a contact route",
"States that a directory listing is consent to be found, not consent to be pitched",
"Flags community scouting as the highest-effort and highest-compliance-cost channel, so it is paired with at least one cheaper channel rather than run alone"
]
},
{
"id": 8,
"prompt": "Rootbound is annual, we're planning edition three, 5 months out, and we've committed internally to this being a long-term thing rather than a one-off. Last year our keynote cancelled 6 days before the event and it nearly sank the day. I start a fresh speaker spreadsheet every year and bin the old one. This time I want to measure the sourcing properly - I was going to track the combined social reach of everyone we confirm, the number of invitations we sent, and average response time. What am I missing?",
"expected_output": "Replaces the reach metric with source effectiveness, names the evidence-completeness threshold, converts the yearly spreadsheet into one running cross-edition roster, and promotes the slow rungs for a compounding mandate with a fresh cancellation.",
"files": [],
"expectations": [
"Names source effectiveness - which channel produced each confirmed speaker - as the metric worth carrying above the others",
"Gives the reason as making the next edition's channel ranking the organizer's own rather than a generic default",
"Rejects combined social reach as a sourcing metric, consistent with excluding audience-size metrics from qualification",
"Names shortlist evidence completeness as the measure carrying a real threshold: every shortlisted name, or the shortlist is not finished",
"Labels the long-list-to-confirmation ratio as self-set and meaningful only against the organizer's own prior editions",
"Instructs the organizer to pick two metrics, write down what each would change next edition, and record both before the event",
"Replaces the yearly fresh-and-binned spreadsheet with one running roster carried across editions",
"Specifies that roster as a categorical role field such as speaker, invited-only, or keynote, plus a free-text detail field",
"States the purpose of that roster as carrying a returning candidate's prior evidence and outcome forward instead of re-qualifying them from zero",
"Promotes organizer reference for headline candidates because last edition's late cancellation is still fresh",
"Promotes community scouting or another slow, compounding channel because the mandate is a long-term annual asset rather than a one-off edition",
"Records which candidates declined and why, alongside which channels produced confirmed speakers"
]
}
],
"trigger_queries": [
{
"query": "Our CFP closes in two weeks and we only have four talks for a ten-slot programme, help me find people to invite",
"should_trigger": true
},
{
"query": "How do I find speakers for our conference outside the call for papers?",
"should_trigger": true
},
{
"query": "We need a keynote on platform engineering who isn't the same three people every European conference books",
"should_trigger": true
},
{
"query": "build me a shortlist of invited speakers for our infra day",
"should_trigger": true
},
{
"query": "we skipped the CFP entirely and the event is in nine weeks, I need a whole lineup",
"should_trigger": true
},
{
"query": "where do I look for speakers from Latin America who can present in English?",
"should_trigger": true
},
{
"query": "our programme has a hole on observability and nobody submitted anything on it",
"should_trigger": true
},
{
"query": "how do I qualify someone I've never seen present before?",
"should_trigger": true
},
{
"query": "I want more women on stage at our devops day, where do I even start",
"should_trigger": true
},
{
"query": "who should we ask to open day one?",
"should_trigger": true
},
{
"query": "how many people should I approach to fill six speaking slots?",
"should_trigger": true
},
{
"query": "we need names for the workshop track, none came through the call",
"should_trigger": true
},
{
"query": "help me put together a list of potential keynotes with the reasoning for each one",
"should_trigger": true
},
{
"query": "is there a directory I can browse by topic and region to find people to invite?",
"should_trigger": true
},
{
"query": "our lineup is all people I already know personally and it shows",
"should_trigger": true
},
{
"query": "how do I check whether someone is actually good on stage before we commit?",
"should_trigger": true
},
{
"query": "we have four invited slots and no idea who to put in them",
"should_trigger": true
},
{
"query": "find me practitioners who have actually run Kubernetes at scale rather than vendors selling it",
"should_trigger": true
},
{
"query": "I need speakers for our user conference, customers ideally, not just our own staff",
"should_trigger": true
},
{
"query": "what's a good way to get speaker referrals out of our past speakers?",
"should_trigger": true
},
{
"query": "our sponsor wants to put their VP on the main stage, how do I handle that",
"should_trigger": true
},
{
"query": "we want first-time speakers on the programme but none of them applied",
"should_trigger": true
},
{
"query": "help me rank these six candidates for our closing talk",
"should_trigger": true
},
{
"query": "how far ahead do I need to lock in someone who's flying in?",
"should_trigger": true
},
{
"query": "the call for proposals didn't fill the schedule and I need a plan B",
"should_trigger": true
},
{
"query": "what should I hand over to the person writing our invitation emails?",
"should_trigger": true
},
{
"query": "can you help me scout people from project forums who'd be good on stage",
"should_trigger": true
},
{
"query": "we run this conf every year and start the speaker hunt from zero each time, that can't be right",
"should_trigger": true
},
{
"query": "our event is regional but we want at least three voices from outside the country",
"should_trigger": true
},
{
"query": "I've got a spreadsheet of 40 possible speakers, how do I narrow it down",
"should_trigger": true
},
{
"query": "should I pay a speaker agency to fill our keynote slots?",
"should_trigger": true
},
{
"query": "how do I decide between two people for the same session?",
"should_trigger": true
},
{
"query": "we need somebody to talk about WASM tooling and I don't know anyone in that space",
"should_trigger": true
},
{
"query": "what evidence do I need before I put a name in front of the programme committee",
"should_trigger": true
},
{
"query": "help me write a brief for the opening slot so I know what I'm looking for",
"should_trigger": true
},
{
"query": "for our second edition I want to track which channels actually produced speakers",
"should_trigger": true
},
{
"query": "I keep finding impressive people who don't fit the programme",
"should_trigger": true
},
{
"query": "how do I find people who've never spoken at a conference before?",
"should_trigger": true
},
{
"query": "vendor conference, 20 sessions, all invited, where do the speakers even come from",
"should_trigger": true
},
{
"query": "we can cover travel but not a fee, does that change who I can approach",
"should_trigger": true
},
{
"query": "how do I avoid inviting the same circle of people again this year",
"should_trigger": true
},
{
"query": "can you help me build a pool of potential speakers that lasts beyond this edition",
"should_trigger": true
},
{
"query": "the programme chair wants three names per open slot by Friday",
"should_trigger": true
},
{
"query": "write the invitation email to the keynote speaker we picked",
"should_trigger": false
},
{
"query": "how many follow-ups before I give up on a speaker who hasn't replied?",
"should_trigger": false
},
{
"query": "our invited speaker asked what our fee is, how do I word the answer",
"should_trigger": false
},
{
"query": "personalize this outreach message for the speaker we shortlisted",
"should_trigger": false
},
{
"query": "what does a first invitation message have to disclose up front?",
"should_trigger": false
},
{
"query": "draft a follow-up to a speaker who went quiet after saying maybe",
"should_trigger": false
},
{
"query": "when should our call for papers open and close?",
"should_trigger": false
},
{
"query": "what fields should the CFP submission form have?",
"should_trigger": false
},
{
"query": "should our call for papers use blind review or named submissions?",
"should_trigger": false
},
{
"query": "how many proposals should one person be allowed to submit to our CFP?",
"should_trigger": false
},
{
"query": "write the speaker benefits section of our call for proposals",
"should_trigger": false
},
{
"query": "what session formats should we offer in the call?",
"should_trigger": false
},
{
"query": "we have 180 CFP submissions, how does the review committee score them?",
"should_trigger": false
},
{
"query": "how do I run conflict-of-interest recusal on our programme committee?",
"should_trigger": false
},
{
"query": "what acceptance rate should we target on our call for papers?",
"should_trigger": false
},
{
"query": "how do I tell 140 people their proposal was rejected",
"should_trigger": false
},
{
"query": "design the scoring rubric our reviewers use on submitted talks",
"should_trigger": false
},
{
"query": "what do I send a speaker after they accept?",
"should_trigger": false
},
{
"query": "how do I run the green room and the tech check on the day?",
"should_trigger": false
},
{
"query": "our speaker needs a hotel booked and a slide deadline, what's the ladder",
"should_trigger": false
},
{
"query": "do we need recording consent from speakers and how does opt-out work",
"should_trigger": false
},
{
"query": "one of our confirmed speakers uses a wheelchair, what does the venue need",
"should_trigger": false
},
{
"query": "should we offer live captioning on all the talks?",
"should_trigger": false
},
{
"query": "do we need a travel grant fund and how do we size it",
"should_trigger": false
},
{
"query": "what should our conference stand for compared to the other three in this city?",
"should_trigger": false
},
{
"query": "do we need a headline name to sell tickets, and what kind of draw",
"should_trigger": false
},
{
"query": "how many talk slots should our two-day event have?",
"should_trigger": false
},
{
"query": "single track or multi track for 400 attendees?",
"should_trigger": false
},
{
"query": "should we run workshops on a separate day or inline with the talks",
"should_trigger": false
},
{
"query": "lay our 24 accepted talks into a two-room grid",
"should_trigger": false
},
{
"query": "how much buffer between sessions in the run of show",
"should_trigger": false
},
{
"query": "write a job description for the conference producer we're hiring",
"should_trigger": false
},
{
"query": "how do I structure the interview loop for an event coordinator role",
"should_trigger": false
},
{
"query": "how do I recruit and brief judges for our hackathon?",
"should_trigger": false
},
{
"query": "what weights should our hackathon scoring criteria have",
"should_trigger": false
},
{
"query": "we need 30 volunteers to staff the event, how do I recruit them",
"should_trigger": false
},
{
"query": "how do I book guests for our developer podcast every week",
"should_trigger": false
},
{
"query": "set up a newsletter swap with another event's organizer",
"should_trigger": false
},
{
"query": "match sponsors with attendees for 1:1 meetings at our event",
"should_trigger": false
},
{
"query": "a government minister is attending, what protocol and escort do we need",
"should_trigger": false
},
{
"query": "source candidates for our open senior backend engineer role",
"should_trigger": false
},
{
"query": "build a prospect list of VP Engineering contacts for our sales outreach",
"should_trigger": false
},
{
"query": "find influencers to promote our product launch",
"should_trigger": false
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |