new-engagement
Skill8 starsChanged 4 months agoMITView on GitHub
Scaffold client engagement projects from the installed project template: parse client, topic, format, and destination; create folders; pre-fill AGENT.md; create a wiki page; initialize git. Trigger: 'new engagement', 'scaffold engagement', '/new-engagement'.
- Deletes or force-pushes
In this repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- demo-project/CLAUDE.mdCLAUDE.md
- project-template/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- close-weekSkill
- end-sessionSkill
- ingestSkill
- lintSkill
- plan-weekSkill
- project-statusSkill
- querySkill
- review-sessionsSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 5 are distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
When the user says "new engagement", "scaffold engagement", or invokes a slash command like /new-engagement, scaffold one or more client …
- Very specific: line 16
If the template is absent but network access is available, the agent may clone https://github.com/ozgurkarahan/ai-agent-memory.git into a …
- Distinctive: line 17
If either root remains unresolved, report the missing root and stop.
- Distinctive: line 18
Remove any temporary clone after scaffolding.
- Very specific: line 20
All wiki paths below are relative to WIKI_ROOT.
- Distinctive: line 33
If the input is ambiguous or missing details, ask clarifying questions:
- Distinctive: line 81
Use the platform-native equivalent on Windows.
- Very specific: line 93
Create the folders with a .gitkeep in each:
- Very specific: line 192
Add the page entry to wiki/projects/_index.md when that index exists, update the Projects section and article count in index.md, and append …
- Distinctive: line 208
After scaffolding all projects, output:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
New Engagement
When the user says "new engagement", "scaffold engagement", or invokes a slash command like /new-engagement, scaffold one or more client engagement projects from the resolved project template.
Resolve required roots
Before scaffolding:
- Resolve
WIKI_ROOTusingmemory/schema.md, thenschema.md, then the memory-wiki path declared inAGENT.md. - Resolve
TEMPLATE_ROOTby checking forproject-template/besideWIKI_ROOT, then the path declared inAGENT.md. - If the template is absent but network access is available, the agent may clone
https://github.com/ozgurkarahan/ai-agent-memory.gitinto a temporary folder and use itsproject-template/directory. - If either root remains unresolved, report the missing root and stop. Do not guess a personal path.
- Remove any temporary clone after scaffolding.
All wiki paths below are relative to WIKI_ROOT.
Step 1: Parse the request
Extract from the user input:
- Client name (e.g., "Acme", "Contoso", "Contoso")
- Topics — one or more engagement topics, each with a format
- Projects root — use an explicit destination from the request or
AGENT.md; otherwise ask where to create the client workspace
Expected input format: <ClientName> — <topic1> (format), <topic2> (format)
Formats: presentation, workshop, demo, presentation + demo, or combinations.
If the input is ambiguous or missing details, ask clarifying questions:
- What is the target audience? (e.g., technical leadership, developers, executives)
- What format? (presentation, workshop, demo, or combination)
- What are the key objectives?
- Any specific technologies or topics to cover?
Step 2: Scaffold each project
For each topic, do the following.
2a. Create project directory
PROJECT_DIR={PROJECTS_ROOT}/{ClientName}/10-projects/{project-slug}
mkdir -p "$PROJECT_DIR"Convert the topic to a kebab-case slug (e.g., "Agent Framework Engagement" → agent-framework-engagement).
The standard client workspace layout is 00-client/ (intel), 10-projects/ (delivery repos), 10-projects/_archive/, 20-assets/ (raw client material), 90-scratch/. Engagements always go under 10-projects/.
2b. Copy template files
Copy the canonical files from TEMPLATE_ROOT. The full canonical scaffold is:
{project}/
├── AGENT.md # MAIN — overview, env, commands, workflow, refs
├── CLAUDE.md # Thin shim → "Read AGENT.md"
├── .claude/
│ ├── CLAUDE.md # Claude Code project config + Project Context block
│ └── commands/status.md # /status slash-command for 30-sec briefing
├── .github/
│ ├── copilot-instructions.md # GitHub Copilot CLI instructions
│ └── instructions/
│ └── end-session.instructions.md # end-session shim (5 steps)
├── .gitignore # tracks .claude/CLAUDE.md, .vscode/; ignores runtime caches
├── LICENSE # MIT
├── README.md # Human-facing
└── (format-specific folders) # slides/, demos/, exercises/, etc.# Copy with the platform's native filesystem tools, then strip template history
cp -R "$TEMPLATE_ROOT/." "$PROJECT_DIR/"
rm -rf "$PROJECT_DIR/.git"Use the platform-native equivalent on Windows. After copying, verify the scaffold matches the canonical structure above; do not copy unrelated files that may have accumulated beside the template.
2c. Create format-specific folders
| Format | Folders to create |
| Presentation | slides/, demos/, docs/ |
| Workshop | slides/, exercises/, solutions/, instructor/, docs/ |
| Demo | demos/, docs/ |
| Presentation + Demo | slides/, demos/, docs/ |
| Mixed/Other | Combine as appropriate |
Create the folders with a .gitkeep in each:
mkdir -p "$PROJECT_DIR/{folder}" && touch "$PROJECT_DIR/{folder}/.gitkeep"2d. Pre-fill AGENT.md
Replace the templated AGENT.md with engagement-specific content:
# {Topic Title}
## Overview
{Client} engagement: {topic description}. Format: {format}. Target audience: {audience}.
## Engagement Details
| Field | Value |
|-------|-------|
| Client | {ClientName} |
| Topic | {Topic} |
| Format | {format} |
| Audience | {audience} |
| Status | Not started |
## Deliverables
{Based on format — see deliverables table below}
## Key Paths
| Path | Description |
|------|-------------|
{format-specific paths}
## Memory Wiki
Root: `{resolved WIKI_ROOT}`
## Reference Documents
| Document | Contents |
|----------|----------|
| `{resolved WIKI_ROOT}/wiki/projects/{project-slug}.md` | Canonical project memory page |
## First Session Instructions
When the agent is first launched in this project, follow these steps:
1. **Read all existing files** in the project to understand current state
2. **Research the topic** — find the latest capabilities, announcements, and best practices for {topic}
3. **Enter plan mode** and propose a detailed engagement plan including:
- Agenda / slide outline with timing
- Key talking points and messaging
- Demo scenarios (if applicable)
- Exercise descriptions (if workshop)
- Competitive positioning where relevant
4. **Wait for user approval** before creating any content files
5. **After approval**, scaffold the content structure and begin developmentDeliverables by format
| Format | Deliverables |
| Presentation | - [ ] Slide outline with talking points<br>- [ ] Speaker notes<br>- [ ] Demo script (if applicable)<br>- [ ] Leave-behind document<br>- [ ] Q&A preparation |
| Workshop | - [ ] Lab guide with prerequisites<br>- [ ] Exercise instructions (progressive difficulty)<br>- [ ] Solution files<br>- [ ] Instructor notes with timing<br>- [ ] Cheat sheet / quick reference<br>- [ ] Prerequisites checklist |
| Demo | - [ ] Demo script with steps<br>- [ ] Setup instructions<br>- [ ] Fallback plan<br>- [ ] Talking points |
2e. Create the project memory page
Project-specific context lives in two places:
AGENT.mdin the project repo — identity, workflow, conventions (single source of truth)wiki/projects/{project-slug}.mdunderWIKI_ROOT— durable project memory (lessons, reference, technical details, related links).
Save the user's engagement request verbatim to raw/data/new-engagement-{project-slug}-{YYYY-MM-DD}.md. If a project memory page does not yet exist, create it with this minimal skeleton:
---
title: {Topic Title}
category: projects
tags: [{client-tag}, {topic-tags}]
source_docs: ["raw/data/new-engagement-{project-slug}-{YYYY-MM-DD}.md"]
date_created: {YYYY-MM-DD}
date_updated: {YYYY-MM-DD}
---
# {Topic Title}
## Summary
## Context
## Project notes
## Open actions
## Related
## Sources
- `raw/data/new-engagement-{project-slug}-{YYYY-MM-DD}.md`Add the page entry to wiki/projects/_index.md when that index exists, update the Projects section and article count in index.md, and append to log.md.
2f. Git init
cd "$PROJECT_DIR" && git init
git add -A
git commit -m "Initial scaffold: {topic} engagement for {Client}
Scaffolded from project-template by the new-engagement skill.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>"Step 3: Output results
After scaffolding all projects, output:
Launch commands
For each project, suggest both agent options. Use whichever CLI the user has installed; the commands are equivalent in scope (allow-all-tools / bypass-permissions). cd first so the agent picks up the project's AGENT.md.
PowerShell (Windows):
cd {actual-project-path}
# GitHub Copilot CLI
copilot --allow-all-tools
# Claude Code (alternative)
claude --dangerously-skip-permissionsBash (Linux / macOS / Git Bash):
cd {actual-project-path}
# GitHub Copilot CLI
copilot --allow-all-tools
# Claude Code (alternative)
claude --dangerously-skip-permissionsSuggested initial prompts
For each project, suggest a specific first prompt based on the topic and format. Pipe it to the agent with -p:
cd {actual-project-path}
copilot --allow-all-tools -p "Research the latest agentic AI capabilities, then propose a presentation outline with demo scenarios for technical leadership"Summary table
| Project | Path | Format | Status |
| {topic} | {actual-project-path} | {format} | Scaffolded |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

