# Marketplace > Free, ready-made folder systems and workflows for Palette Desktop. Pick one, install it, and let your agent adapt it to your work. Canonical URL: [https://palette.team/marketplace](https://palette.team/marketplace) Palette Desktop: [https://dl.palette.team/desktop/stable/Palette.Desktop_universal.dmg](https://dl.palette.team/desktop/stable/Palette.Desktop_universal.dmg) # Company OS Canonical URL: [https://palette.team/marketplace/company-os](https://palette.team/marketplace/company-os) Your whole company as a file system: one folder of Markdown for how the company works and what is active across accounts, GTM, product, engineering, operations, and people. ## At a glance - Type: Folder system - Best for: Businesses and teams that want one shared folder for company context that people and agents can use. - Key outputs: Company-wide shared context, Team operating spaces, Customer and teammate records, Decision and memory logs - Version: 1.0.4 - Install files: 130 - Requires: Palette Desktop 0.5+ - Maintained by: Palette - License: MIT Company OS gives businesses that want to work smarter with AI one shared folder for how the company works. It installs spaces for accounts, GTM, product, engineering, operations, and people, plus shared instructions and memory, so supported agents start each session with the same company context. ## What folder structure does Company OS install? ```tree company-os/ # One shared home for how the company works ├── README.md # The company index and starting point ├── accounts/ # Customer context and relationship health ├── gtm/ # Brand, market, playbooks, and active work ├── product/ # Strategy, discovery, specs, and feedback ├── engineering/ # Architecture, decisions, and runbooks ├── ops/ # Admin, finance, hiring, and investor work ├── people/ # How to work with each teammate ├── .agents/commands/ # Onboarding and upkeep workflows ├── lessons.md # What agents should do differently next time └── log.md # A concise history of meaningful work ``` - **Six connected spaces** give every team a clear home while keeping company context readable across the whole system. - **Shared agent instructions** make different AI tools follow the same company rules. - **Lessons and decision logs** keep corrections, decisions, and useful history available in future sessions. - **Built-in workflows** add customers, teammates, and new spaces without hand-copying folders. ## How do you bring in the context you already have? Say "set me up," then start with the best source you already have: - Use a connected source such as Palette's Context Library to bring in company, team, and active-work context. - Drop existing decks, wiki exports, org charts, and other documents into `_inbox/`. - Use a built-in copy-and-paste prompt to bring useful context over from ChatGPT, Claude, Notion, or another tool. - Answer one short round of questions if the context is not written down yet. Your agent sorts what it finds into the right spaces, flags what is still missing, and confirms the result before replacing the template. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How does Company OS work with different AI agents? Company OS keeps your context in ordinary Markdown files that Claude Code, Codex, Gemini, and other supported agents can read. Shared instructions tell each agent where to look and how to work, while the linked spaces give it the company context relevant to the task. You can move from product work to a launch, a customer question, or an operations task without rebuilding the company brief each time. New decisions and durable lessons stay in the folder for the next person and the next agent. ## What can you ask once Company OS is set up? Once the company context is in place, any supported agent can work from it in normal language: - "Help me bring the company context I already have in ChatGPT into this workspace." - "Review this launch using our brand, product, and customer context." - "What should I know from across the company before working on our hiring plan?" ## Common questions ### What does Company OS organize? Company OS organizes accounts, GTM, product, engineering, operations, and people in one shared folder, with instructions and memory files that supported agents can read. ### Does Company OS arrive filled with my company context? It starts as a blank folder structure. During setup, an agent uses available company context and asks for anything missing before replacing the placeholders with your information. --- # GTM OS Canonical URL: [https://palette.team/marketplace/gtm-os](https://palette.team/marketplace/gtm-os) Your go-to-market as a file system: brand, market, strategy, playbooks, accounts, team context, and active work in one folder of Markdown that your team and AI agents can read. ## At a glance - Type: Folder system - Best for: Founders and GTM teams coordinating sales, marketing, and customer work. - Key outputs: GTM foundations, Campaign and content workspaces, Account context, Shared GTM memory - Version: 1.1.4 - Install files: 71 - Requires: Palette Desktop 0.5+ - Maintained by: Palette - License: MIT GTM OS gives founders and commercial teams one shared folder for how they sell, market, and work with customers. It installs brand, market, strategy, accounts, playbooks, team context, and active work, so people and supported agents can reuse the same go-to-market context in every session. ## What folder structure does GTM OS install? ```tree gtm-os/ # One shared home for your go-to-market ├── README.md # The GTM index and starting point ├── foundations/ # Stable context behind commercial decisions │ ├── brand/ # Messaging, positioning, and voice │ ├── market/ # Audience and market context │ ├── product-marketing/ # Product truth and sales narrative │ └── strategy/ # The go-to-market strategy ├── accounts/ # Customer briefs and relationship health ├── playbooks/ # How the team approaches recurring GTM work ├── templates/ # Reusable formats for launches and campaigns ├── team/ # Teammate profiles and working styles ├── work/ # Campaigns, projects, content, and research in flight ├── .agents/commands/ # Onboarding, account, review, and upkeep workflows ├── lessons.md # What agents should do differently next time └── log.md # A concise history of meaningful work ``` - **Stable foundations** keep brand, market, product, and strategy close to the work they shape. - **Playbooks and templates** turn the team's best methods into reusable starting points. - **Shared instructions, lessons, and history** help different agents work from the same rules and past decisions. - **Built-in workflows** cover onboarding, accounts, teammates, reviews, and upkeep. ## How do you bring in the GTM context you already have? Say "set me up," then give GTM OS the best context you already have: - Use a connected source such as Palette's Context Library to bring in company, customer, team, and active-work context. - Drop positioning documents, sales decks, research, account exports, and other material into `_inbox/`. - Use a built-in copy-and-paste prompt to bring useful context over from ChatGPT, Claude, Notion, or another tool. - Answer one short round of questions about your company, market, brand, and team. Your agent sorts that material into foundations, accounts, playbooks, team context, and active work, then confirms any gaps before replacing the template. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How does GTM OS work with different AI agents? GTM OS keeps your brand, market, product story, accounts, playbooks, and active work in ordinary Markdown files that Claude Code, Codex, Gemini, and other supported agents can read. Shared instructions help each agent work from the same GTM context instead of relying on a separate prompt for each tool. The same context supports sales, marketing, and customer work. One agent can prepare an account call, another can draft a launch, and both can use the same positioning, voice, customer history, and team decisions. ## What can you ask once GTM OS is set up? Once the GTM context is in place, any supported agent can use it for the sales, marketing, or customer work in front of you: - "Help me bring our positioning, brand voice, and campaign history over from Notion." - "Prepare me for the Acme call using our account history and sales approach." - "Turn this product update into launch messaging for our audience and voice." ## Common questions ### What folder structure does GTM OS install? GTM OS installs foundations, active work, accounts, playbooks, templates, team context, and memory logs in one go-to-market workspace. ### How does GTM OS work with different agents? A shared AGENTS.md contains the main instructions, with thin pointers for supported agents so they follow the same workspace rules and context. --- # Product OS Canonical URL: [https://palette.team/marketplace/product-os](https://palette.team/marketplace/product-os) Your product work as a file system: user feedback, pain points, feature requests, discovery, strategy, specs, competitors, and team input in one folder of Markdown. ## At a glance - Type: Folder system - Best for: Product teams turning user signal into evidence-backed decisions and specs. - Key outputs: Linked product evidence, Feedback dashboard, Discovery and strategy records, Paste-ready product specs - Version: 1.0.6 - Install files: 73 - Requires: Palette Desktop 0.5+ - Maintained by: Palette - License: MIT Product OS gives product teams a clear path from raw user signal to a decision. It installs connected homes for feedback, interviews, requests, pain points, discovery, strategy, specs, competitors, and team input, so agents can synthesize evidence while keeping source links visible and projects in the issue tracker. ## What folder structure does Product OS install? ```tree product-os/ # One shared home for product work ├── README.md # The product index and starting point ├── user-feedback/ # Raw feedback and the patterns found across it │ ├── feedback/ # Short user comments and observations │ ├── user-interviews/ # Longer customer conversations │ ├── feature-requests/ # Explicit asks │ ├── painpoints/ # Recurring pain points linked to evidence │ └── index.html # A filterable feedback dashboard ├── discovery/ # Open questions and learning plans ├── strategy/ # Durable claims about how the product wins ├── specs/ # Paste-ready product briefs ├── competitors/ # Capability-focused competitor profiles ├── team-input/ # Roadmap input and dated team recaps ├── documentation/ # Product overview and shared terminology └── .agents/commands/ # Feedback, synthesis, spec, and upkeep workflows ``` - **Linked product evidence** keeps raw feedback, interviews, requests, and pain points traceable instead of becoming a pile of quotes. - **A feedback dashboard** makes signal easy to filter by company, type, sentiment, and tag. - **A path from signal to decision** connects discovery, strategy, and specs to their sources. - **Built-in workflows** capture feedback, find patterns, draft specs, and review the system. ## How do you bring in the product context you already have? Say "set me up," then start with the material your product team already uses: - Use a connected source such as Palette's Context Library to bring in company, team, and active-work context. - Drop product briefs, research exports, feedback folders, interview notes, and existing strategy documents into `_inbox/`. - Use a built-in copy-and-paste prompt to bring useful context over from ChatGPT, Claude, Notion, or another tool. - Answer one short round of questions about the product, users, team, and feedback sources. Your agent sorts that material into the right product folders, keeps the source links visible, and confirms gaps before replacing the template. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How does Product OS work with different AI agents? Product OS keeps feedback, research, discovery, strategy, specs, competitors, and team input in ordinary Markdown files that Claude Code, Codex, Gemini, and other supported agents can read. Shared instructions help each agent work from the same product evidence instead of starting from a summary trapped in one chat. That means you can switch agents without rebuilding the product brief. An agent can trace a question back to raw feedback, connect it to existing strategy, and help with the next piece of product work while your issue tracker remains the home for projects and tasks. ## What can you ask once Product OS is set up? Once the product context is in place, any supported agent can work from the same evidence, strategy, and past decisions: - "Help me bring our product brief, research, and feedback into this workspace." - "What are customers repeatedly saying about onboarding, and what evidence supports it?" - "Draft a product spec grounded in our feedback, discovery, and strategy." ## Common questions ### What folder structure does Product OS install? Product OS installs user feedback, discovery, strategy, specs, competitors, team input, product documentation, shared instructions, and memory logs in one workspace. ### Does Product OS replace my roadmap or issue tracker? No. Initiatives, projects, and tasks stay in your issue tracker. Product OS keeps the evidence, discovery, strategy, and paste-ready specs behind those decisions. --- # Project OS Canonical URL: [https://palette.team/marketplace/project-os](https://palette.team/marketplace/project-os) A single project as a file system: one folder for context, meetings, decisions, research, design, deliverables, ways of working, and a generated status dashboard. ## At a glance - Type: Folder system - Best for: Agencies, consultancies, and internal teams running one shared project. - Key outputs: Project brief and goals, Meeting and decision records, Deliverable tracking, Generated status dashboard - Version: 1.0.4 - Install files: 55 - Requires: Palette Desktop 0.5+ - Maintained by: Palette - License: MIT Project OS gives agencies, consultancies, and internal teams one shared folder for a single project. It installs the brief, goals, meetings, decisions, research, deliverables, team context, and a generated status dashboard, so the team, the client, and supported agents can work from the same current project picture. ## What folder structure does Project OS install? ```tree project-os/ # One shared home for a single project ├── README.md # The project index and starting point ├── overview.html # A status dashboard generated from the files ├── context/ # Brief, goals, stakeholders, and shared language ├── meetings/ # Dated notes and follow-ups ├── decisions/ # Numbered decisions with their reasoning ├── research/ # Findings and interview notes ├── design/ # Design references and working links ├── deliverables/ # Milestones and project outputs ├── team/ # Onboarding and ways of working ├── admin/ # Scope and timeline └── .agents/commands/ # Onboarding, meeting, decision, and upkeep workflows ``` - **One project folder** keeps the information people and agents need before they act. - **A status dashboard** shows goals, stakeholders, milestones, and recent activity without maintaining another presentation. - **Decision and meeting workflows** keep important context structured and easy to revisit. - **Lessons and project history** carry useful context into the next session. ## How do you bring in the project context you already have? Project OS includes a sample project so you can see the finished structure first. When you say "set me up," you can replace it using the context you already have: - Use a connected source such as Palette's Context Library to bring in the project, people, and status. - Drop the brief, statement of work, kickoff notes, meeting exports, and other documents into `_inbox/`. - Use a built-in copy-and-paste prompt to bring useful context over from ChatGPT, Claude, Notion, or another tool. - Answer one short round of questions if important context is still in people's heads. Your agent sorts the material, confirms any gaps, replaces the sample, and updates the dashboard. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How does Project OS work with different AI agents? Project OS keeps the brief, goals, stakeholders, meetings, decisions, research, deliverables, and status in ordinary files that Claude Code, Codex, Gemini, and other supported agents can read. Shared instructions help each agent work from the same current project picture instead of depending on one person's chat history. That makes handoffs easier across the team and across agents. One agent can prepare a client meeting, another can review a deliverable, and both can see the same scope, decisions, recent work, and open questions. ## What can you ask once Project OS is set up? Once the project is set up, any supported agent can work from the same brief, decisions, and recent history: - "Help me bring the brief, meeting notes, and project history over from our current tools." - "Prepare me for the client sync using the brief, recent meetings, and open work." - "What changed this week, and which risks or decisions need attention next?" ## Common questions ### What folder structure does Project OS install? Project OS installs project context, meetings, decisions, research, design, deliverables, team context, administration, shared instructions, and memory logs. ### How does the Project OS dashboard stay current? The included overview.html reads project information from the workspace files, so the status view can be regenerated from the same context your team and agents maintain. --- # Accounts Canonical URL: [https://palette.team/marketplace/accounts](https://palette.team/marketplace/accounts) A lightweight account system for sales and CS. Track the companies you're engaged with in plain files, log meetings consistently, keep a running health read, and prep for calls using your own sales approach. ## At a glance - Type: Workflow - Best for: Sales and customer success teams managing active relationships. - Key outputs: Account briefs, Meeting notes and follow-ups, Relationship health reads, Pre-call briefs - Version: 1.0.2 - Install files: 5 - Requires: Current Palette Desktop release - Maintained by: Palette - License: MIT Accounts gives sales and customer success teams working memory for active relationships. It installs account briefs, meeting notes, health reads, and a roster next to your own sales approach, so your team and its agents can prepare calls and follow-ups from the context that never fits neatly into a CRM field. ## What does the Accounts workflow install? ```tree your-folder/ # The folder where you add the workflow └── accounts/ # Working memory for active customer relationships ├── accounts.md # The roster, owners, stages, and next steps ├── approach.md # Your sales and discovery approach ├── _template/ # The starting point for every new account └── / # One folder per customer or prospect ├── README.md # Current context and account strategy ├── health.md # Risks, momentum, and recommended attention ├── meetings/ # Dated notes and follow-ups └── shareables/ # Material prepared for the customer ``` - **One account brief** keeps the current situation and strategy easy to scan. - **A shared roster** shows owners, stages, last touch, and next steps across accounts. - **Your own sales approach** shapes call prep and discovery instead of a generic playbook. ## What does Accounts look like in practice? ### Prepare for a customer call - **You ask:** "Prep me for the Acme call." - **The agent works from:** The account brief, recent meeting notes, the health read, and your sales approach. - **You get:** A pre-call brief with the relationship stage, key people, priorities, tailored discovery questions, risks, and a suggested next step. ### Log a customer meeting - **You ask:** "Log these notes from my Acme check-in." - **The agent works from:** The notes you provide and the existing account brief. - **You get:** A dated meeting file, refreshed open threads and history, plus an updated last touch and next step in the account roster. ## How does Accounts adapt to your sales process? 1. Install Accounts from Marketplace in Palette Desktop. 2. Your agent finds the right home for the account folder and adapts the sales approach to your company, product, and customers. 3. Add an account, log a meeting, or ask for a call brief. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## What can you do after installing Accounts? ```commands Create an account for Acme Corp. # start the first account Log these notes from my Acme meeting. # capture notes and follow-ups Prepare me for the next Acme call. # build a grounded call brief Refresh Acme's relationship health. # update risks and momentum ``` ## Common questions ### What does the Accounts workflow install? It adds an accounts workspace with reusable account and health templates, a shared roster, your sales approach, and a workflow for creating accounts, logging meetings, preparing calls, and refreshing relationship health. ### How does Accounts prepare me for a call? When you ask for call preparation, the workflow reads the account brief, health read, meeting history, next steps, and your sales approach to produce a grounded brief. --- # Content Canonical URL: [https://palette.team/marketplace/content](https://palette.team/marketplace/content) A content system for writers and marketers: start drafts for a target channel, refine them against your own playbook, check copy against your brand voice, plan distribution, and track what you ship. ## At a glance - Type: Workflow - Best for: Writers and marketing teams producing content across several channels. - Key outputs: Channel-aware drafts, Brand voice checks, Distribution plans, Published content log - Version: 1.0.2 - Install files: 5 - Requires: Current Palette Desktop release - Maintained by: Palette - License: MIT Content gives writers and marketing teams one home for the work they publish. It installs channel-aware drafts, a content playbook, a personalized voice guide, distribution planning, and a shipped-content log, so an agent can help turn rough ideas into finished work while staying grounded in how your team actually writes. ## What does the Content workflow install? ```tree your-folder/ # The folder where you add the workflow └── content/ # One home for content work ├── content.md # Drafts, scheduled work, and published pieces ├── playbook.md # How your team approaches content ├── voice.md # Your tone, language, and copy self-check └── drafts/ # Work in progress ├── README.md # The structure every draft follows └── .md # One draft with its channel and key takeaway ``` - **A personalized voice guide** gives every draft the same tone and language rules. - **A content playbook** keeps writing and distribution grounded in how your team works. - **One workflow** helps start, refine, voice-check, distribute, and log a piece. ## What does Content look like in practice? ### Turn an idea into a channel-aware draft - **You ask:** "Help me write a LinkedIn post about why we changed our onboarding." - **The agent works from:** Your voice guide, content playbook, the target channel, the one takeaway, and relevant company or product context. - **You get:** A dated draft shaped for LinkedIn and a new Draft entry in the content log. ### Check whether a draft sounds like you - **You ask:** "Does this sound like us?" - **The agent works from:** The draft and your personalized voice guide. - **You get:** Specific lines that drift from your voice, the rule each line breaks, and a suggested fix instead of a generic rewrite. ## How does Content adapt to your brand voice? 1. Install Content from Marketplace in Palette Desktop. 2. Your agent adapts the voice guide from the context it can find and asks for anything missing. 3. Start with an idea, choose a channel, and refine the draft against your real voice. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## What can you do after installing Content? ```commands Help me write a post about why we rebuilt onboarding. # start a draft Does this draft sound like us? # check the voice Plan how we should distribute this piece. # prepare distribution Show me our current and published content. # review the content log ``` ## Common questions ### What does the Content workflow install? It adds a content workspace, a reusable draft structure, a content playbook, a personalized voice file, and one workflow for drafting, refining, checking, distributing, and logging content. ### How does Content learn my brand voice? During installation, the agent uses available brand context and asks for anything missing before personalizing the voice file that future voice checks read. --- # Competitor Analysis Canonical URL: [https://palette.team/marketplace/competitor-analysis](https://palette.team/marketplace/competitor-analysis) A structured competitor-tracking workflow with a market landscape, consistent competitor profiles, sourced research, and a repeatable way to add and update competitors. ## At a glance - Type: Workflow - Best for: GTM, product marketing, and strategy teams tracking a changing market. - Key outputs: Market landscape, Competitor profiles, Threat assessment, Sourced research - Version: 1.0.2 - Install files: 2 - Requires: Current Palette Desktop release - Maintained by: Palette - License: MIT Competitor Analysis gives founders, GTM, product marketing, and strategy teams a repeatable way to understand a changing market. It installs a market landscape, consistent competitor profiles, sourced research, and update workflows, so each new profile adds to the wider market view. ## What does the Competitor Analysis workflow install? ```tree your-folder/ # The folder where you add the workflow └── / # Your chosen home for market research ├── landscape.md # Market tiers, threats, and open questions └── competitors/ # One consistent profile per competitor ├── .md # Positioning, pricing, risks, and dated sources ├── .md # Another company in the same profile format └── .md # Add more as the market changes ``` - **Consistent profiles** make competitors easier to compare. - **The market landscape** brings tiers, threats, and open questions together. - **Research workflows** add a competitor and update the wider market view in one pass. ## What does Competitor Analysis look like in practice? ### Add a competitor to the landscape - **You ask:** "Add Glean to our competitor landscape." - **The agent works from:** Current public sources, your product context, the existing competitor format, and the market landscape. - **You get:** A dated, sourced competitor profile plus an updated tier table and threat assessment where warranted. ### Refresh an existing profile - **You ask:** "Refresh Acme's profile before our positioning review." - **The agent works from:** The existing profile, current product and pricing information, recent news, customer discussions, and your positioning. - **You get:** An updated profile with current sources, plus a refreshed landscape if the competitor's place or threat has changed. ## How does Competitor Analysis set up your market? 1. Install Competitor Analysis from Marketplace in Palette Desktop. 2. Choose the market folder and the first competitors to track. 3. Let your agent research them now, or leave clear placeholders to fill later. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How do you keep competitor research current? ```commands Add Acme to our competitor landscape. # research and add a company Compare Acme and Globex using our current research. # use the shared format What changed in our market landscape? # synthesize the profiles ``` ## Common questions ### What does the Competitor Analysis workflow install? It adds a market landscape, a consistent profile format, and workflows for researching a company, adding its profile, and updating the wider market view. ### Can it start with competitors I already track? Yes. During installation, the agent asks which competitors to seed and whether to research them immediately or leave structured placeholders for later. --- # Presentations Canonical URL: [https://palette.team/marketplace/presentations](https://palette.team/marketplace/presentations) A presentation system for any role: draft sales pitches, product reviews, team updates, or conference talks in Markdown, then generate self-contained HTML slide decks with presenter notes and PDF export. ## At a glance - Type: Workflow - Best for: Anyone creating repeatable sales, product, team, or conference presentations. - Key outputs: Markdown presentation drafts, Self-contained HTML decks, Presenter notes, PDF-ready slides - Version: 1.0.2 - Install files: 5 - Requires: Current Palette Desktop release - Maintained by: Palette - License: MIT Presentations gives anyone making sales, product, team, or conference decks a repeatable path from rough idea to finished presentation. It installs story-first Markdown drafts, reusable branded slides, and a self-contained HTML deck engine with presenter notes and PDF export, so you can shape the message before designing every slide. ## What does the Presentations workflow install? ```tree your-folder/ # The folder where you add the workflow └── presentations/ # One home for presentation work ├── presentations.md # A record of completed decks ├── snippets.md # Reusable company, team, and closing slides ├── drafts/ # Outlines and slide notes │ └── .md # The key message and slide-by-slide story └── slides/ # Browser-ready presentations ├── template.html # The reusable deck design and controls └── .html # A generated, self-contained deck ``` - **Story-first drafts** keep the message clear before design takes over. - **Browser-ready decks** include keyboard navigation, fullscreen, speaker notes, a timer, and PDF export. - **Personalized slide snippets** make recurring company and team slides easy to reuse. ## What does Presentations look like in practice? ### Shape the story before making slides - **You ask:** "Help me build a 20-minute Q3 pipeline review for the leadership team." - **The agent works from:** The audience, event details, duration, your company context, voice, and the presentation draft structure. - **You get:** A dated Markdown draft with one key message, a timed outline, and slide-by-slide notes. ### Turn the approved story into a deck - **You ask:** "Turn the Q3 pipeline review into slides." - **The agent works from:** The approved draft, the HTML deck template, and your personalized company and team slide snippets. - **You get:** A self-contained HTML deck with minimal slides, speaker notes, presenter controls, and browser-based PDF export. ## How does Presentations adapt to you? 1. Install Presentations from Marketplace in Palette Desktop. 2. Your agent personalizes the reusable slides from the context it can find. 3. Draft the story, generate the deck, and open the finished HTML file in a browser. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## How do you create a deck after installation? ```commands Help me shape a Q3 pipeline review. # start with the story Turn the approved draft into slides. # generate the browser-ready deck Log the presentation I gave today. # update the presentation history ``` ## Common questions ### What does the Presentations workflow create? It creates structured Markdown drafts and self-contained HTML slide decks with keyboard navigation, themes, presenter notes, a timer, and browser-based PDF export. ### Does Presentations require a slide-building service? No. The generated deck is a self-contained HTML file with no build step or external dependency, so you can open it directly in a browser. --- # Workspace Heal Canonical URL: [https://palette.team/marketplace/workspace-heal](https://palette.team/marketplace/workspace-heal) A workspace maintenance workflow that checks broken links, stale indexes, unfinished setup, and instruction drift. It reports what is wrong, what is safe to fix, and what needs your judgment without imposing a folder structure. ## At a glance - Type: Workflow - Best for: Teams maintaining file-based workspaces used by people and AI agents. - Key outputs: Workspace health report, Broken-link findings, Instruction-drift findings, Safe-fix plan - Version: 1.0.3 - Install files: 1 - Requires: Current Palette Desktop release - Maintained by: Palette - License: MIT Workspace Heal checks file-based workspaces for broken links, stale indexes, unfinished setup, and instruction drift. It starts with a read-only report, follows the rules already present in the folder, and separates safe repairs from changes that need your judgment. ## What does Workspace Heal check? ```tree your-workspace/ # The folder Workspace Heal reviews ├── **/*.md # Local links and stale paths ├── **/README.md # Indexes compared with files on disk ├── AGENTS.md # Instructions checked for drift ├── SETUP.md # Active setup markers checked ├── generated files # Rebuild guidance checked when declared └── other folders # Reviewed against the workspace's own rules ``` - **A clear health report** groups findings by impact and includes exact file paths. - **Workspace-aware checks** follow the folder's own guidance instead of imposing a new structure. - **A safe-fix split** shows what can be repaired conservatively and what needs review. - **Coverage notes** explain what was checked, excluded, or left incomplete. ## What does Workspace Heal look like in practice? ### Check a workspace before making changes - **You ask:** "Check this workspace's health." - **The agent works from:** The workspace README, agent instructions, local documentation indexes, links, setup markers, and any conventions the folder declares. - **You get:** A health summary, the review scope, findings grouped by impact, and a clear split between safe fixes and decisions that need you. ### Repair the safe structural issues - **You ask:** "Fix the safe structural issues." - **The agent works from:** The health report and the workspace's own rules for links, indexes, generated files, and instructions. - **You get:** Confirmed repairs followed by a fresh check. Nothing is moved, renamed, or deleted without your approval. ## How do you start a Workspace Heal check? 1. Install Workspace Heal from Marketplace in Palette Desktop. 2. Open the folder you want to review. No special setup or folder structure is required. 3. Ask your agent to check the workspace's health. Palette Desktop creates a checkpoint before installation, so you can restore the folder if you want to start over. ## What can you ask after installing Workspace Heal? ```commands Check this workspace's health. # start with a read-only report Find broken links and stale indexes. # focus the review Fix the safe structural issues. # apply conservative repairs Why is this folder hard to navigate? # find structural friction ``` ## Common questions ### What does Workspace Heal check? It checks local Markdown links, indexes, setup markers, generated-file guidance, workspace instructions, naming rules, and other conventions the workspace actually declares. ### Does Workspace Heal change files automatically? No. It starts with a read-only report and only applies safe structural fixes when you explicitly ask. It asks before moving, renaming, or deleting files, or making content decisions.