sales / skill
AI Marketing Panel
A synthetic customer-panel kit with evidence-based composite dossiers, separate draft reactions, synthesis and scoring workflows.
At a glance
- Purpose
- A synthetic customer-panel kit with evidence-based composite dossiers, separate draft reactions, synthesis and scoring workflows.
- Runs in
- Chat app with skills, Local agent with shell
- Operating systems
- macOS, Linux
- Needs
- Runtime: Node 22.18.0+, bash and coreutils (shell mode); Needs internet: Not stated; Needs accounts or API keys beyond your agent: Not stated
- Files and data
- Reads files: Yes; Writes files: Yes; Sends data to other services: Not stated
- Tool type
- Skill
- Delivery
- From GitHub
- Version
- Latest on the main branch
- Access
- Code on GitHub
- Compatibility
- According to the README, its install table marks the Claude app (drag and drop the kit zip into the conversation) and Codex (unzip the kit to a folder, then point Codex at that folder) as working on 2026-08-12. It marks Claude on the web as presumed to work the same way but not yet tried. The README describes full runs as Claude Code today and chat-only runs as Claude on the web or ChatGPT. The README says the room lifecycle needs bash and coreutils, while panel validators require Node 22.18.0 or later. It mentions macOS and Linux in its Node install advice and says its room script parses on Apple's stock bash 3.2.
- External costs
- The repository is MIT-licensed. The README says a run uses model usage that varies with the model, the number of active seats and the length of the artifact, and that its own cost comparison figures are unmeasured estimates. The README says hosted install and tuning are available, without giving a price for them. Your coding agent or model provider may charge for usage.
- Permissions
- The skill acts through your coding agent with the access you give it. The README asks users to create owned kit copies and work in their generated workspaces. The README describes a bash room-lifecycle script, room resets and archiving, and Node tools for initializing and validating panels and preparing and validating runs. The README's Node upgrade advice includes a Homebrew installation on macOS or distribution packages on Linux. According to the README, panel work uses customer evidence such as support tickets, sales calls and reviews.
- Usage terms
- MIT License, as stated in the repository. See its LICENSE file for the exact terms.
Code on GitHub
rodaddy/ai-marketing-panel
MITSynthetic customer panel: test marketing copy against a calibrated panel of AI personas before it ships. MIT.
Public GitHub data fetched · 0 days old
Skill folders in this repository (1)
Folders that contain a SKILL.md file, from GitHub's file list on .
What it does
AI Marketing Panel is a kit for running a synthetic customer panel. According to the README, it starts with a decision about a landing page, price or email and builds customer composites from evidence such as support tickets, sales calls and reviews. The README says each composite receives a sourced dossier and reacts to the draft in a separate context.The README describes separate reactions, a synthesis preserving disagreements, a scored verdict, rewrites with retests when useful, and records of predictions and later outcomes for calibration.
The README lists the synthetic-customer-panel skill, workflows, standards, scripts, workspace templates and installation documents. The README calls the panel a structured directional signal rather than a substitute for real customers, and tells users to make an owned copy for each use.
Who it is for
Owners and small teams who write their own marketing copy and have nobody to show it to before it ships.Example tasks
Build a panel of five to fifteen customer composites from support tickets, sales calls and reviews.Run each panel seat on its own against one landing page, price or email draft.
Read a synthesis of the reactions that keeps the disagreements.
Score a draft with owner-approved weights and ask for three rewrites.
Record predictions and later actual outcomes for calibration.
Getting started
The README's install table lists two routes: drag and drop the kit zip into a Claude app conversation, or unzip the kit to a folder and point Codex at it. It names INSTALL.md as the canonical agent-executed installation contract and START-HERE.md as the first handoff file. To check the Node floor it gives bash skills/synthetic-customer-panel/scripts/check-node.sh. The README has the exact commands.Limitations
According to the README, the base kit should be copied for each use rather than run in place. The README describes T1 as using subagents and workflow orchestration in Claude Code today. The README says shell-only runs use separate fresh contexts sequentially, losing speed and some cross-seat isolation strength. The room lifecycle needs bash and coreutils; panel validators require Node 22.18.0 or later. The README says older Node versions leave panel and run validation on the self-report path.The README says chat-only use turns scripts into manual checklists, makes script-enforced gates advisory and requires manual archiving. The README labels runs degraded when seats share a context. The README calls the panel directional evidence that does not replace real customer contact, legal review, accessibility testing, usability testing, analytics or market evidence. The repository can change at any time; check its README before you install.