Library
42 features in 10 categories. Each has a Read page and a skill. Every card starts with a plain-words description and when you would use it; the technical summary is underneath.
Foundation9
Bootstrap Operating Behavior
The ground rules for how the setup assistant behaves while it works with you. It looks at what you already have before it asks anything, adds to your existing instructions instead of wiping them out, and stays out of your way.
Use it when you want setup to feel like working with a helpful colleague instead of filling in a form. It won't ask questions it can answer itself, and it won't overwrite work you've already done.
Technical summary
How the bootstrap reads state, merges and enriches instead of overwriting or merely preserving, acknowledges pointer-style features, and routes unresolved gaps
Agent Personality & Tone
Gives your Claude project a steady voice: warm, friendly, patient and honest, with no made-up character. It also sets a sensible default for situations your instructions don't cover.
Use it in almost every project, especially if people who are new to AI assistants will be using it. They get a patient helper instead of a terse or confusing one.
Technical summary
Generically warm, patient tone grounded in integrity, organization, and helpfulness — no named personality, no personal contact
Process Safeguard Standing Rule
A simple safety valve. If one task goes around in circles for three or more back-and-forths without getting anywhere, Claude stops and steers you back to your real work.
Use it to keep small things from eating a whole session, like endlessly tweaking a color, a profile field or a preview.
Technical summary
Redirects toward real work if any single task stalls 3+ exchanges
Agent Role & Project Identity
Tells Claude what this project is and what its job is in it: the project's name and description, what kind of project it is, and a short code used to label documents. If your description is vague, it sharpens it instead of just copying it.
Use it first in any new project. Everything else builds on it, and it is what makes Claude act like a specialist for your work instead of a generic helper.
Technical summary
Project title, description, type, project code, and a project-appropriate agent role — sharpened rather than merely copied forward when existing content reads generic or thin
Connectors & Capabilities
Keeps a written list of the tools Claude is connected to in this project (file access, email, Microsoft 365 and so on) and what each one lets it do. Each new conversation checks the list and warns you if something has dropped off.
Use it whenever Claude needs to work with your files or accounts, so you learn about a broken connection at the start instead of halfway through a task.
Technical summary
Records which connectors are live and what each enables
File Storage & Transfer Rules
Sets the basic rules for where files live and how they move: what stays in Claude's Project Files and what stays on your computer, the one official project folder, a drop folder for incoming files, and the rule that files are handed to you as ordinary downloads, never as scrambled blocks of text.
Use it before Claude touches any files. It prevents lost, duplicated or garbled files and gives you one obvious place to drop things for Claude to pick up.
Technical summary
Two storage layers, canonical path, #drop\, and the no-base64 transfer rule
Multi-Sub-Project Awareness
For a folder that really holds several separate projects under one umbrella, such as one client with different matters. Each sub-project gets its own self-contained folder, shared material sits at the top, and there is a rule for where dropped files go.
Use it when one project folder is really several projects and mixing them would cause confusion.
Technical summary
Formalizes a project root containing multiple genuinely separate sub-projects — sub-project table, self-contained per-sub-project folders, shared root-level layer, drop-folder routing, and file-naming extension; extends FTR-07/FTR-08 rather than replacing either
Email Account Awareness & Signature Management
Records which email account this project sends from when you have several set up. It also captures your signature and writing style from an email you really sent, never an invented one, and saves it in two places so it is applied the same way every time.
Use it if Claude drafts or sends emails for you and you want them to come from the right account and sound like you. It needs an email connection already set up on your computer.
Technical summary
Project-level email account selection plus a deterministic signature (and default composition style) pulled from a real sent email, dual-written into admin\EMAIL_SIGNATURE.md and inline in admin\AGENT_INSTRUCTIONS.md
Project Cache
Keeps the files Claude reads at the start of every chat up to date. One staging folder holds the current copies, along with an automatically kept index that tells Claude which file to open and when. To refresh, you drag that one folder into the project.
Use it to give Claude dependable background on a project without re-explaining it each time, and to make sure the changes you make to your files actually reach Claude. It is one of the most useful pieces for working with AI agents.
Technical summary
Foundation context contract: one cache folder (project_cache\) plus one generated index (INDEX.md, backed by INDEX.json) that tells any agent which cached file to read and when; flat staging dir with a manifest (Excel via FTR-36 when installed, otherwise markdown), explicit permanent exclusion for admin\AGENT_INSTRUCTIONS.md and CLAUDE.md, which never belong here regardless of how they turn up in a scan
Organization2
Folder Structure
A standard folder layout for your project (a drop folder, documents, drafts, research, brand, trackers, archive and more), plus extra folders that fit the type of project you have. It only creates what is missing.
Use it to start a project with a tidy, predictable filing system, so you and Claude always know where things go and where to find them.
Technical summary
Baseline project folder layout plus type-specific additions
File Naming, Track IDs & Versioning
A consistent way to name and number your files so every document can be pointed to without confusion: one standard file name format, short document ID numbers grouped by topic, a version number on the file name and another inside the document, and a method for rolling several documents up into one summary.
Use it once a project will have more than a handful of documents. Instead of 'the March letter, second draft', you get a clear ID, and you stop hunting for which copy is current.
Technical summary
Naming format, multi-track document IDs, dual versioning, roll-up pattern
Documents5
Document Workflow & Status System
A simple life cycle for documents (Needed, Seeded, Draft, Active, Final, Restricted) with plain commands for creating and saving them. It also warns you when something was mentioned in conversation but never actually written down.
Use it to keep track of what is finished and what is still just an idea, and to make sure good information from a chat doesn't get lost.
Technical summary
Document lifecycle, seeded-item warning, and the command set
PDF Rule
One firm rule: PDFs are for sharing finished work, not for making it. The real document is always the Word file, the PDF is just the printout, and Claude never edits a PDF directly.
Use it to avoid the classic mess of an edited PDF that no longer matches the original, so your master file is always the one that is right.
Technical summary
PDFs are publication-only; never edit a PDF directly
Markdown Sidecars
When a PDF or similar file turns up, Claude can write a plain-text companion file that captures everything in it, page by page, including descriptions of images and signatures. It can also pull the contents out of Excel files the same way.
Use it when you have important PDFs that Claude keeps needing to re-read. Once the text copy exists, the content is searchable and quick to use without reopening the original.
Technical summary
On-demand faithful .md companions from PDFs; markdown preference; now includes a tested, approved .xlsx extraction method (raw zip/XML parse, UTF-8 explicit)
Directory Cleanup & Inventory
A thorough, read-only survey of a big folder. Claude reads every file, notes which ones belong together, writes a text companion for each readable file, sorts everything by what it is actually about (not by its file name), suggests a better folder layout and produces one master report. It never moves or deletes anything.
Use it on a large, messy or unfamiliar folder, like an inherited drive or an old project, when you need to understand what is there before deciding what to do.
Technical summary
Full recursive scan, sidecars, categorization, folder recommendations, master report — no moves/deletes
Paste-Driven Provenance Archive
For sources that only you can open, like paywalled sites or systems you have to log in to. You copy each item and paste it into the chat. Claude saves each one in a standard format that shows where it came from and keeps a running index. Claude never fetches or scrapes anything on its own.
Use it to build a trustworthy reference library from material behind a login or paywall, with a clear record of where every piece came from.
Technical summary
A person retrieves items from a source an agent can't reach (paywalled, session-gated, access-controlled) and pastes them in one at a time; the agent files each into a fixed provenance template and a running index — never fetches or scrapes on its own
Trackers3
Document Registry
A spreadsheet that lists every document in the project, with its status, in one master list. Claude updates it in a careful, checked way and keeps a short guide to how it works.
Use it when you have a lot of files and want one place to see what exists, what state it is in and what comes next, like an air-traffic board for your documents.
Technical summary
Excel-workbook tracker (via FTR-36), single source of truth, commit protocol, README, index
Active Flags Tracker
A short spreadsheet of things that are stuck or waiting: open blockers, unfinished action items and answers you are waiting on from someone else. Each one gets a permanent number, and they are shown at the start of every session.
Use it so nothing stays stuck quietly. A document can be finished and still have an open question attached to it, and this keeps that question in front of you.
Technical summary
Two-sheet Excel workbook (via FTR-36) — open blockers and pending items surfaced at session open; new stable FLAG-NN ID column
Excel Tracker (Base Capability)
The groundwork for using Excel spreadsheets as trackers, with one row per item. It sets the safe way for Claude to create and edit them (using PowerShell and Excel on your PC), makes a backup before every edit and checks the result in five steps.
Use it when you track things you want to sort, filter and check off yourself. Several other trackers in this library are built on it. It needs a Windows PC with Excel.
Technical summary
PowerShell + Excel COM automation as the only approved access method (genuinely fast in practice, the real reason it's default), pre-edit backup convention, five-step verification discipline, priming-vs-transfer guidance, and a project-cache markdown sidecar convention — the base capability domain-specific Excel trackers (e.g. legal discovery) build on
Research4
Multi-AI Research Synthesis System
A method for research questions that are too important or too tricky to trust to one quick answer. Claude writes a careful research prompt and shows it to you. You run it on several AI services and drop the results in your drop folder, and Claude combines them into one polished master document.
Use it for decisions that matter, like comparing options, checking facts or digging into a complicated topic, where you want several AIs to cross-check each other.
Technical summary
RS track: write a research prompt, save it to a single consistent folder, present it inline in chat for the person to copy, submit to multiple AI platforms (with friendly guardrails on making sure each submission is a genuine deep-research run, not just a normal answer), polish the results into one authoritative filed master document
Cascading Research Pipeline
Runs several linked research questions in order, in one sitting, for when later questions depend on earlier answers. It pauses so each answer is actually loaded before it moves on to the questions that depend on it.
Use it for bigger strategic research where you can't ask the right second question until you have seen the first answer, such as 'what are our options?' followed by 'which one fits our situation?'
Technical summary
Orchestrates FTR-28 across a dependency graph of research steps — parent chain ID with numbered sub-steps, cascading and parallel dependencies, a pause-and-confirm gate so a step's output is actually loaded into context before anything depending on it runs, and a full worked example
Cowork Orchestrated Research
A hands-off way to run a batch of research. You give Cowork one instruction covering many research questions, run side by side or one after another, and it works through them and reports as each one finishes. No pasting each one by hand.
Use it when you have a long list of research items and don't want to babysit every one.
Technical summary
A third execution mode alongside Deep (FTR-28) and Cascade (FTR-30) — one kickoff instruction hands Cowork a whole batch of research items (parallel or chained), and Cowork self-sequences and reports per-item as results land, instead of manual per-item submission
Prompt Vetting Checklist
A quick five-question review of a research request before you send it: is the framing right, does it cover your goals, are the questions specific enough, do the cross-references hold up, and is anything out of date. It runs only when you ask.
Use it before sending an important research prompt to several AIs, so you don't spend time and money on a weak question.
Technical summary
A structured, five-question review a research prompt goes through before it's considered genuinely ready to submit — role & framing, goal coverage, objective granularity, cross-reference correctness, staleness/currency flagging. A quality gate between drafting a prompt and running it through FTR-28's pipeline, on request only, not automatic
Brand & Identity2
Brand Setup
Captures your look once: logo, colors and fonts. It tries out a few styles with you and then writes a Brand Standards document that every document Claude makes afterward follows.
Use it when you want letters, reports and forms that all look like they came from the same business, without describing your style again every time.
Technical summary
Logo, color, typography, preview, and the Brand Standards document
Entity / Organization Profile
Records the basic facts about the business or organization behind the project (legal name, address, tax numbers, mission, key contacts) one time, so Claude can reuse them everywhere.
Use it when you often fill out applications, proposals, contracts or forms, so you stop retyping the same details. Skip it for purely personal projects.
Technical summary
One-time org/business profile referenced on every document
Legal5
Legal Preferences Profile
Captures how your legal documents should look: caption style, formatting, file naming, how you cite cases and how motions are laid out. It includes a ready-made generic motion template to start from.
Use it on a legal project so every filing Claude drafts matches your court's expected format from the first draft. It isn't needed outside legal work.
Technical summary
Caption, formatting, naming, citation standards for legal projects — now packaged with a generic Motion Branding Spec and a downloadable generic motion template
Legal Citation Management System
Tracks every case or statute you plan to cite in one report, with a status on each: still to be verified, verified, flagged or dropped. The aim is to catch AI-invented or misdescribed citations before they reach a filing. Nothing goes out below 'verified'.
Use it for any court filing or legal research where Claude helps draft. AI can make up or mischaracterize cases, and this is your safeguard against that.
Technical summary
Citation Vetting Report (CVR) tracking every legal citation through TBV/VERIFIED/FLAG/DROP status — catches hallucinated or mischaracterized AI-generated citations before filing
Legal Case Research Synthesis
The legal version of the research method. Research that could strengthen an argument or settle a question about local court practice is tracked separately and tied to your case's master list of open items.
Use it in a litigation project when you need careful research that feeds directly into your filings or case decisions.
Technical summary
LR track: FTR-28's engine, tied to the Master Case File instead of a generic tracker, with legal-specific case-context requirements, RS/LR track-assessment guidance, and a note on how FTR-30's chain mechanic applies to the LR track too
Georgia Appellate Rules & Statutes Reference
A reference library with the current, source-checked text of the Georgia Court of Appeals rules, the Georgia Supreme Court rules and the state's appeals statutes (O.C.G.A. Title 5), with a download package and a list of recent changes.
Use it when preparing or checking a Georgia appeal, so Claude works from the actual rule text instead of memory. It is reference material only and not legal advice.
Technical summary
Complete, primary-source-verified current text of the GA Court of Appeals Rules, GA Supreme Court Rules, and O.C.G.A. Title 5 (Appeal and Error); ships with a downloadable reference package plus a changelog — companion to FTR-37, not overlapping
Legal Discovery Tracking
A spreadsheet tracker for the discovery questions in a lawsuit (interrogatories, requests for production and requests for admission), with one row per question and columns for status, response and outcome. It includes Georgia's rules for counting questions against the limit and a guide to organizing strategy documents.
Use it when you are serving or answering discovery in a Georgia case and want to see every request, its status and its outcome at a glance.
Technical summary
Excel tracker schema for Interrogatories/RFPs/RFAs — deterministic row-per-question tracking, Georgia discovery-cap counting rules, party-folder structure, and a 7-category strategy-document taxonomy; ships with a downloadable Georgia statutes-and-guide reference package
Utility2
EML Archiver (Base Capability)
Collects emails as exact original copies (.eml files) and files them in a permanent, dated, labeled archive with an index of what was saved. It works with Microsoft 365 by script, with a manual browser method as a fallback, and comes with a downloadable toolkit.
Use it when you need provable, unaltered email records, for a dispute, an audit or a handover.
Technical summary
Retrieve email as byte-authentic .eml files (Office 365 Graph API, scripted, or manual browser reconciliation fallback) and file into a permanent, labeled, chronological archive with sidecars and a manifest; ships with a downloadable toolkit (scripts + reconciliation HTML example)
Email Topic Discovery
Finds every email about a topic, from a person, or matching a search term, across one or more mailboxes, and hands them to the email archiver to be saved properly.
Use it when you need to answer 'what did we actually send and receive about this?', such as in a dispute or a staff transition. It needs the EML Archiver installed.
Technical summary
Topic-scoped email discovery: define a scope (search term/domain/person), search/capture/reconcile/fetch/archive/describe pipeline nesting under FTR-39's retrieval mechanics; ships with a downloadable run-descriptor template
Continuity5
Startup Display & Reminder Mode
A snapshot shown at the start of every conversation: where the project lives, what is waiting in the drop folder, which trackers are active, open flags, brand status and which tools are connected. You can switch it off.
Use it if you want to pick a project back up at a glance, without having to ask 'where were we?' every time.
Technical summary
Opening project-snapshot dashboard; toggleable on/off
Session Handoff Protocol
When a conversation gets too long, Claude writes a summary document that captures everything important. You paste it into a fresh conversation and carry on where you left off.
Use it when a chat is running out of room, or when you want to move to a new conversation without losing the thread.
Technical summary
Carry a conversation into the next one without losing context
Cowork Integration
A routine for handing a focused job to Cowork (Claude's mode for carrying out longer tasks on your files) and bringing the result back. Claude writes the starting instructions, you run them there, and the output lands in your drop folder to be filed. Claude will also suggest this itself for big, repetitive jobs.
Use it for large mechanical jobs, like working through a pile of files, where each item needs little judgment but there are a lot of them.
Technical summary
Offload a task to Cowork and bring the result back into #drop; now proactively suggests Cowork for high-volume, mechanical, low-judgment-per-item work instead of waiting to be asked
CLAUDE.md Cowork Context File
Makes a short file called CLAUDE.md at the top of your project folder, which Cowork and Claude Code read automatically. It is a guide that points them to your rules and procedures, not a copy of your instructions, so they load only what they need. Your full instructions stay whole in the original file.
Use it if you also work with Cowork or Claude Code in this folder. When you add a new rule or feature later, you change the instructions only; CLAUDE.md is updated only when you approve a re-run.
Technical summary
Root CLAUDE.md as a short generated entry file, not a mirror: it routes agents to rules, procedures and reference sliced from the agent instructions, which stay whole and are the source of truth. Adding a new rule, procedure or feature edits the instructions only; CLAUDE.md changes only when the owner approves a re-run of the conversion
Repository Ledger Visualization
On request, makes a good-looking one-page picture of your whole project folder, with files grouped into a few meaningful categories and counted. It uses the same saved design every time.
Use it when you want a quick, presentable overview of what a project contains. Just say 'show me the project ledger'. It never appears on its own.
Technical summary
On-demand, on-brand HTML snapshot of the whole file repository — folders/files grouped into a small number of meaningful categories, rendered from a one-time-derived, locked design template rather than redesigned per request; the first feature in this catalog to use the POINTER-STYLE mechanism, with its raw HTML/CSS template and a fabricated-data live preview hosted as their own standalone asset files
Health & Maintenance5
Project Audit / Health Check
Say 'run an audit' and get a color-coded, one-page health check of your project: what is solid, what needs attention and what is broken, each with a suggested fix.
Use it every so often, or whenever something feels off, to spot problems early and get a clear to-do list.
Technical summary
On-demand color-coded health-check matrix with recommended actions
Agent Instructions Sync
Your project instructions live in two places: Claude's Project Instructions box and a file on your computer. At the start of a conversation Claude compares the two and tells you if they have drifted apart.
Use it so you never end up working from an out-of-date copy of your own rules.
Technical summary
Keeps the Project Instructions field and the local file aligned
Agent Instructions Generation & Deploy
The final wrap-up step. It assembles the finished instructions file from everything you have set up, checks it and puts it in place. It also adds the standard housekeeping sections: open items, key documents, ongoing duties, limitations, how to rerun setup and a change log.
Use it last, after the other features you want are installed, to turn all your choices into one working instructions file.
Technical summary
Assembles, verifies, and deploys the instructions + housekeeping sections, enriching rather than merely preserving flagged content
Install Manifest & Reconciliation
Keeps a list in your instructions of which catalog features you have installed and at which versions. On request it checks that list against the current catalog and tells you what is up to date, what is out of date and what you haven't installed. It never changes anything by itself.
Use it to find out whether your project is behind on improvements, and then decide for yourself what to update.
Technical summary
Records what's installed and at what version; on-demand check for updates against this catalog, plus a deeper opt-in mode that compares actual implemented content against each feature's real spec page, not just version numbers
Preferred MCP Instance
If your project has more than one connection of the same kind (for example, several tools that control your computer), this lets you name the one this project should use. Claude doesn't have to guess, and two chats don't bump into each other.
Use it if you run several Claude conversations at the same time and they get in each other's way.
Technical summary
Names which MCP connector instance this project should default to when more than one of the same kind is installed, so parallel conversations don't collide — a plain preference, not a management system; commands: "show me the MCPs" and "prefer [name]"