samber/dev-event-organizer-skills / skills/event-planning-timeline
event-planning-timeline
VerifiedBuild the months-long work-back plan for a tech event, any edition - the doors-open anchor date, parallel tracks (venue, program/CFP, sponsors, marketing, ticketing) each worked back from that date, cross-track dependency gates, a buffer policy through latest-safe commitment dates, and replanning checkpoints that cut scope instead of moving the date. Use whenever asked to build an event planning timeline, conference retroplanning, a work-back or reverse schedule, event milestone sequencing, or to replan an event running behind schedule. Scheduling mechanics only. Do NOT use to pick the date itself - use samber/dev-event-organizer-skills@event-date-selection - or for the day-of run of show - use samber/dev-event-organizer-skills@event-run-of-show.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 21.1 KB |
| evals/evals.json | config | 24.3 KB |
| references/buffers-and-replanning.md | markdown | 7.9 KB |
| references/tracks-and-gates.md | markdown | 5.5 KB |
references/buffers-and-replanning.md
View on GitHub# Buffers, checkpoints, and the mid-flight replan
Depth behind SKILL.md's buffer-policy and replanning-trigger menus, plus a worked illustrative example. Everything sourced is labeled. Every date in the worked example is invented to show the pattern.
## The two sourced buffer mechanics, kept distinct
Both come from a major community-conference organizing guide's cost-control advice. They are two different tools, not one "have contingency" idea:
**Delayed commitment** shrinks a single line item late.
- The guide's own words: don't overcommit on headcount, don't pre-pay the venue until you must, and order food late - "it's always easier to add a few extra plates as opposed to having too much food ordered."
- The mechanic generalizes to any commitment where the two directions of error are asymmetric. Under-committing early and scaling up late is cheap (extra plates, extra seats, a second swag run). Over-committing early and scaling down late is expensive or impossible (a deposit, a placed order, a signed contractor).
- On the plan, this becomes a _latest-safe date_ per irreversible commitment, and the discipline of not committing before it just because the money is available.
**The scope-tier ladder** re-scopes the whole event at a checkpoint. Before selling a ticket, define three costed versions:
- **Good**: bare-minimum viable, the floor below which the event doesn't run.
- **Better**: the expected niceties.
- **Best**: real headroom, the extra evening activity, the nicer badges.
Then write dated downgrade checkpoints driven by revenue actually received: on each date, if revenue hasn't reached the tier's threshold, the event downgrades one tier - a decision that fires automatically instead of being argued mid-crisis. Sourced rationale: for sponsorship-funded community events, this decouples "will we run at all" from "did tickets sell," so a soft sales signal downgrades the event instead of cancelling it.
Milestone padding, the third rung, has one legitimate use: lead times outside the team's control (visas, printing, shipping, venue paperwork). Everywhere else it fails silently - everyone plans to the padded date, so the buffer is spent before anyone decides to spend it.
## Checkpoint design
A checkpoint is four fields, all written in advance:
| Field | What it holds |
| ------------------ | ------------------------------------------------------------------------------------------------------------------------ |
| Date | A moment already meaningful in the plan - an announcement, early-bird close, a tier's revenue deadline, one month out |
| Condition | One measurable check: revenue vs. tier threshold, registrations vs. expectation, a gate's prerequisite done or not |
| Pre-written action | Exactly what fires on failure: cut this scope item, downgrade one tier, escalate this push, trigger the replan procedure |
| Owner | Who executes the action without a meeting |
The mechanic generalizes the sourced go/no-go downgrade dates above. The specific conditions and thresholds are always self-set per event - this file deliberately gives none. For demand thresholds at sales moments, defer to `samber/dev-event-organizer-skills@event-market-fit` rather than inventing numbers.
A negative example, the checkpoint that isn't one: "Mid-March: review ticket sales." - no condition, no action, no owner. Mid-March arrives, sales are soft, the review notes it, and the actual decision happens six weeks later as an emergency. The pre-written action is the entire difference between a checkpoint and a status meeting.
## The mid-flight replan procedure
When invoked on an event already behind (interview Q6), or when a checkpoint fires the replan action:
1. Rebuild the actual state per track: which milestones are done, late, or not started - against the plan, not against memory.
2. Recompute each track's remaining runway to its terminal milestone and flag every gate now at risk.
3. Spend buffers explicitly and in order: first un-spent latest-safe slack (commitments that can still slide), then a scope-tier downgrade if the ladder exists, then scope cuts.
4. Cut scope at the gate level: cut the item whose removal un-blocks the most downstream milestones, not the item that is merely most behind.
5. Never move an announced date. Before announcement, a date move is a legitimate lever - price it against the external deadlines (venue windows, sponsor fiscal years) before using it.
6. Re-write the remaining checkpoints against the new plan: a replan that keeps the old checkpoint conditions re-fires them immediately.
7. Tell the user which lever was spent and what it bought - a silent replan hides exactly the information the debrief needs.
## Worked illustrative example
**Illustrative of the general pattern only.** Every date and offset below is invented for a fictional mid-size recurring community conference: none of it is a sourced runway. For the sourced, format-driven runways an actual plan should be calibrated against, use `samber/dev-event-organizer-skills@event-first-edition`.
Anchor: doors-open October 15.
External immovables:
- Venue holds its date only until January 31 (contract deadline).
- Two anchor sponsors close their budgets in April.
| Track | Working backward from October 15 (illustrative dates) |
| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Venue & logistics | Walkthrough Oct 1 ← build-up plan Sep 1 ← deposit **latest-safe Jun 30** ← contract Jan 31 (external immovable) ← locked Jan 15 |
| Program | Announced Jun 15 ← speakers confirmed May 30 ← selection May 15 ← CFP closes Apr 30 ← CFP opens Feb 15 ← designed Feb 1 |
| Sponsors | Fulfillment scheduled Jul 1 ← agreements Jun 1 ← outreach starts Mar 1 (pulled early by sponsors' April budget close) ← prospectus Feb 15 ← tiers costed Feb 1 |
| Ticketing | Final headcount **latest-safe Sep 20** (caterer's real lead time) ← early-bird closes Jul 15 ← sales open Apr 1 ← payment infra live Mar 15 |
| Marketing | Final push Oct 1 ← program-driven push Jun 15 (gated on program announced) ← date announced Apr 1 (gated on payment infra + venue contract) ← channels ready Mar 1 |
Gates visible in the table:
- Date announced Apr 1 waits on payment infra (Mar 15) and the venue contract (Jan 31).
- The sponsor track is pulled two months earlier than internal logic would place it, by an external fiscal-year deadline.
- Two latest-safe dates (venue deposit, final headcount) carry the delayed-commitment buffer.
Checkpoints for the same fictional event (conditions and thresholds invented, as they always are - self-set per event):
| Date | Condition | Pre-written action | Owner |
| ------ | ------------------------------------------------------ | --------------------------------------------------------------------------- | --------------- |
| Apr 30 | Anchor-sponsor revenue below the Better-tier threshold | Downgrade Best → Better; cut the evening event from the plan | Finance owner |
| Jul 15 | Early-bird sales below expectation at close | Escalate the community outreach push; hold catering at Good-tier quantities | Marketing owner |
| Sep 15 | Any gate into the final month still unmet | Fire the replan procedure above | Lead organizer |
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 |