Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
日本語の概要は準備中です。原文の説明を表示しています。
Use for event agency operations: client intake, project scoping, vendor management, crew coordination, production planning, pre-show logistics, and post-show wrap. Delegates technical execution to godmode-eventtech and bdbmediastorm.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
The agency-ops layer for Hybridlabor Global — the business and production-coordination side of running an event company. This skill owns everything that happens before, around, and after the show. It does not own the technical execution itself.
Techniques adapted from talkvalue/event-agency-skills (Apache-2.0). Prose only — no upstream code, no scripts, no dependencies.
| Task | Route to |
|---|---|
| Signal flow, protocol binding, OSC/DMX, hardware limits, MCP orchestration | godmode-eventtech |
| Live show architecture brainstorming, creative direction | bdbmediastorm |
| Lighting cue execution, patch, playback | bdb-grandma3-mcp |
| Video clip, layer, and output control | bdb-resolume-mcp |
| 3D scene, render, spatial build | godmode-3d-creation |
| Client intake, scoping, budget, scheduling, vendor, crew, pre-show logistics, post-show wrap | stay in this skill |
Boundary rule: if the question is "what should this look like" or "what signal goes where", route out. If it is "who is bringing it, what did they promise, and what happens when they don't", it stays here. This skill never issues a technical execution instruction — it routes.
An event request is not a booking until these are known. Anything guessed here becomes a change order later.
| Field | Why it is required | If missing |
|---|---|---|
| Date (and whether date or time is flexible) | Drives every downstream deadline in this skill | Cannot scope. Ask first. |
| Venue (named, or shortlist with constraints) | Venue holds all vendor access; determines power, load-in window, noise curfew | Cannot plan load-in or vendor sequencing |
| Brief — what must happen, who must be in the room | Distinguishes an event from a venue rental | Ask: "What does success look like when the last guest leaves?" |
| Budget range — a band, not a number | Drives the action policy: which classes are never vs criteria | Do not quote. Scope verbally only. |
| Technical rider | AV, staging, power, network, FOH position | Route to godmode-eventtech for feasibility |
| Attendance estimate | Drives catering counts, staffing, registration, capacity | Cannot close catering or staffing |
| Client contact + decision authority | Who can actually say yes | Every later "we'll confirm" is a no |
Produce a written scope before any deposit. It must name: date, venue, attendance band, the ten action classes with their modes (default-deny — see .aos/factory.yaml), what is explicitly out of scope, and the change-order trigger. A scope without an out-of-scope list will be expanded by the client by default.
Run this at the start of a working day on an active event, before a production meeting, or after time away. It produces a point-in-time snapshot, not a live feed.
| Type | Key signals | Default tier |
|---|---|---|
| Vendor | AV, catering, security, decor, transport, staffing, rentals, print | 2 |
| Client | Contracting organisation, corporate domain, C-suite/VP | 2 |
| Sponsor | Activation budget or in-kind contribution — distinct from the client even when the client also sponsors | 3 |
| Speaker | Bureau domain, rider, session, keynote, bio/headshot request | 3 |
| Venue | Hotel, convention centre, outdoor site, venue coordinator, venue-employed catering manager | 2 |
| Internal | Own domain, team aliases, automated notifications | 3 |
When a thread is ambiguous, default to the type with the higher production impact: Vendor > Venue > Speaker for operational threads.
| Tier | Label | Response window | Criteria |
|---|---|---|---|
| 1 | Immediate | 1–2 hours | Blocked deliverable, payment deadline within 48h, client waiting on confirmation, venue or vendor escalation, load-in unresolved within 72h of the event |
| 2 | Today | Business hours | Advancing information requests, assets for review, draft approvals, speaker logistics more than 72h out |
| 3 | Tracking | None | FYIs, confirmations of receipt, vendor acknowledgements, threads you are CC'd on |
These override the table above:
Rationale: the tier table is a prior. The overrides exist because the failure mode of event email is a request that looked like an FYI and was not one.
A dated digest containing: event name, timestamp, event phase, period covered, counts per tier, then the tier-1 and tier-2 threads with the specific action each needs and who owns it. Every extracted action item must be one of:
Never mix these three. An action with no owner is not an action item.
Track existing vendor commitments and deliverables. This skill does not source vendors, negotiate contracts, or process payments.
| Type | Typical lead time | Failure impact |
|---|---|---|
| AV / Technical | 2–4 weeks | Critical — no AV, no event |
| Venue | Contracted months ahead | Critical — venue holds all vendor access |
| Catering / F&B | 2–3 weeks (final count 72h) | High — dietary and count changes cascade |
| Transport / Logistics | 1–2 weeks | High — gear and people movement |
| Security | 1–2 weeks | High — compliance and safety |
| Entertainment / Talent | Rider-dependent | High — contracted, non-fungible |
| Decor / Floral | 1–2 weeks | Medium — visual, not operational |
| Staffing Agencies | 1–2 weeks | Medium — replaceable if flagged early |
| Signage / Print | 5–10 business days | Medium — wayfinding and branding |
| Rentals | 1 week | Medium — tables, chairs, linens |
Urgency is relative to event proximity, never absolute time. The same 48-hour silence is fine in month one and a production risk in advance week.
| Phase | Window | Overdue threshold | Escalation speed |
|---|---|---|---|
| Normal | T-30 and out | 48h no response | Standard: email → wait → follow-up |
| Planning | T-30 to T-8 | 24h no response | Accelerated: email → same-day follow-up |
| Advance | T-7 to T-2 | 12h no response | Urgent: email → phone within 4h |
| Load-in | T-1 | 2h no response | Emergency: phone immediately, escalate to production lead |
| Show day | T-0 | 1h no response | Emergency: phone + contingency activation |
| Status | Meaning |
|---|---|
| OVERDUE | Past deadline, no confirmation |
| AT RISK | Deadline approaching, last contact exceeds the phase threshold |
| ON TRACK | Confirmed, or within the normal response window |
| BLOCKED | Waiting on us, not the vendor |
BLOCKED items are the most important row in the table and are routinely under-reported. When a vendor is waiting on our headcount, floor plan, or approval, that is our team's problem. Surface it first.
| Attempt | AV / Venue / Catering | Decor / Signage / Rentals | Staffing / Transport |
|---|---|---|---|
| 1st | Follow-up email quoting the original request | Follow-up email | Follow-up email |
| 2nd | Phone the account manager within 4h | Phone next business day | Phone next business day |
| 3rd | Escalate to production lead + contingency research | Source a backup vendor | Source a backup vendor |
On show day: skip email entirely. Phone or in-person only. Never send a follow-up email on T-0.
| Age | Status | Action |
|---|---|---|
| Within terms | Current | None |
| 1–30 days | Nudge | Friendly reminder — assume oversight |
| 31–60 days | Firm request | Direct ask with date and amount |
| 60+ days | Escalation | Account lead; consider late fee or hold |
Variance alerts: Venue 5% over (largest single line, small % = big dollars), Catering 10% (per-head counts fluctuate), Marketing 15% (most flexible), Contingency 0% drawn before T-14 (should not be touched), all others 10%.
Site survey — non-negotiable for any first-time venue. Record: power distribution and available amperage by area, network drops and their throughput, load-in vehicle access and dock dimensions, door widths for the largest scenic element, FOH position and sightlines, noise curfew, and rigging points with their load ratings. Route the technical read to godmode-eventtech — this skill records the findings and flags blockers, it does not validate the signal plan.
Paperwork gate. Nothing loads in without: certificate of insurance (COI) from every vendor on site, venue access permits, and crew credentials. These are the most common cause of a T-1 access failure.
Run-of-show (ROS) is the single coordination artefact. One document, versioned, owned by the production lead, distributed to every vendor. It defines: load-in window and sequence, soundcheck, doors, show start, each cue block, and load-out. A conflict discovered in the ROS is cheap. A conflict discovered at the dock is not.
Speaker materials — deadline framework:
| Milestone | Due | Deliverables |
|---|---|---|
| D-30 | Bio (25 / 50 / 100-word variants, third person) + headshot (min 300×300px) | |
| D-21 | Session abstract (75 words, attendee-focused), learning outcomes, AV requirements | |
| D-14 | Final slide deck, event template applied | |
| D-7 | Travel, hotel, ground transport, dietary and accessibility requirements | |
| D-1 | Attendance and schedule confirmed |
Learning outcomes use action verbs only — implement, apply, build, create, deploy, evaluate, identify, master. Never: explore, discuss, learn about, understand, discover, dive into, unpack.
Banned in any bio or session description: thought leader, visionary, guru, passionate about, world-class, cutting-edge, revolutionary, groundbreaking, leverage, synergize, transformative, innovative. Replace with specific numbers, named clients, measurable outcomes, concrete credentials.
| Phase | Anchor | What must be true to advance |
|---|---|---|
| Load-in | T-1 or earlier | COI and permits verified per vendor, power patched and tested, ROS sequence confirmed with the dock |
| Soundcheck | Fixed slot, never "when ready" | Line check done, cue stack loaded, FOH and monitors positioned, walk-through with the client if contracted |
| Doors | Published start | House open, accessibility routes clear, front-of-house briefed on the ROS |
| Show | Per cue block | Production lead confirms each block against the ROS; deviations logged, not improvised |
| Strike | After clearance | Equipment accounted for against the load-in manifest, rentals checked out before the truck leaves |
The strike manifest is the load-in manifest. Count on the way in, count on the way out. Rentals left on site are invoiced at loss.
Within 72 hours of strike:
Performance tiers: Exceeded (beat goal by 10%+ on primary metrics), Met (within 10%), Missed (more than 10% below on one or more), Mixed (hit some, missed others — say which and why).
Report structure: performance summary → what worked (with evidence) → what did not (with likely cause) → key insights (each tied to a data point) → specific recommendations for the next event → sponsor ROI if applicable.
Channel attribution is inherently imperfect — most attendees meet several touchpoints before registering. Report it honestly, acknowledge mixed sources, and never over-credit a single channel.
godmode-eventtech / bdbmediastorm rather than answered from this skill.まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.
日本語の概要は準備中です。原文の説明を表示しています。
Use when a hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).
日本語の概要は準備中です。原文の説明を表示しています。
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.
日本語の概要は準備中です。原文の説明を表示しています。
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.
日本語の概要は準備中です。原文の説明を表示しています。
Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.
日本語の概要は準備中です。原文の説明を表示しています。
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.
日本語の概要は準備中です。原文の説明を表示しています。