littlebird-toolkit

Quality assurance pass for anything shipping under the user's name. Trigger on 'does this sound like me', 'check this draft', 'run this past my voice', 'does this read as AI', 'fix the tone on this', 'my teammate wrote this, make it sound like me', 'de-AI this', or when the user pastes a draft and asks whether it is publishable. Runs a supplied draft, theirs or a teammate's or one an AI produced, against their installed personal voice skill. Returns a marked-up draft with every flagged span, a counted tell inventory, a clean rewrite in their voice, a separate fact-check pass on claims it could not corroborate against their own record, and a short why-each-change section. Tone-corrects only and never invents facts. Does not promise to defeat AI detectors, because nobody honestly can.

Per-client health rollup for agencies, consultancies and freelancers. Trigger on "which client is about to churn", "client health check", "am I losing this client", "weekly client review", "is this account at risk", "scope creep on this client", "which client is eating my margin", "client went quiet". Builds a ranked risk view per client from meetings, message threads and captured dashboards: unmet promises in both directions, silence gaps against that client's own cadence baseline, scope creep with the transcript quote where the ask happened, room and register changes, and payment or renewal signals. Bands with named evidence, never a score. Internal view only, nothing goes to the client.

The full-pipeline personal voice skill builder for Littlebird users. Combines BOTH data sources: mines the user's Littlebird memory via the Littlebird MCP (captured writing, messages, meeting speech, background context) AND ingests their Facebook data export (posts, messages, profile info, with a screenshot-guided export walkthrough). Produces one calibrated voice skill that writes exactly like the user and reads 100% human. Use this skill whenever a user says "build my voice skill", "make Claude write like me", "clone my voice", "create my writing style skill", or wants the most complete version of their voice from every source available. If the user only has ONE source, use facebook-voice-creator or littlebird-voice-creator instead.

New hand-raisers into the CRM, daily. Trigger on 'who engaged with my post today', 'pipe my new comments into the CRM', 'add today's leads to GoHighLevel', 'anyone new since yesterday', 'get these commenters into my CRM', 'daily lead drip', 'tag my new leads by campaign'. Watches the user's own posts for new public engagement since the last run, dedupes each person against the CRM, upserts or produces a correctly shaped import file, tags by campaign, and queues a drafted first message. Drafts only. Never sends, and never writes private message contents into a CRM.

Commitment ledger from Littlebird meetings. Trigger on "what did I promise", "what do people owe me", "who owes me what", "did I follow up", "open action items", "what fell through the cracks", "weekly commitment review", "chase my follow-ups". Harvests the Action Items, Decisions and Open Questions already in every Littlebird meeting summary, checks each one against downstream screen and message evidence to see whether it actually got done, ages what is still open, and drafts nudges for approval. Produces a two-column ledger of what you owe and what you are owed, plus a decisions log.

Competitive intelligence, competitor tracking, market radar, weekly competitive digest, competitor deep dive, new entrant detection, watchlist. Logs every sighting of a tracked competitor or market shift that crossed your screen (posts, screenshots, demos, community threads, calls), ranks by velocity rather than volume, catches names entering your field of view for the first time, and reconciles that against external research on pricing, positioning, launches, funding, and personnel.

Turn one long-form artifact into a week of voice-matched derivatives that each take a different angle. Trigger on 'repurpose this post', 'turn this into a week of content', 'make content from this call', 'break this down for LinkedIn', 'I need a content pack', 'what else can I do with this webinar'. Produces 5 to 7 standalone short posts, a carousel outline with per-slide copy, an email, a thread, a sequencing plan with reasons, and a mandatory do-not-publish list. Enforces a seven-angle taxonomy so the pack is genuine variety and not the same idea five times. Drafts through the user's own voice skill when one is installed. Never publishes and never posts.

Morning brief, daily digest, what's on today, brief me on my day, start my day, daily rundown, what do I need to know today, what changed since yesterday, what should I do first. Builds one screen covering today's schedule with a reason each meeting matters, the commitments actually due, what went cold, the genuinely important unread threads, one highest-leverage action with its reasoning shown, and what changed since yesterday. Rolls up sibling routine reports rather than re-deriving them. Runs as a daily routine, or on demand. Requires the Littlebird MCP.

What did I do today, write my dev log, what did I work on last night, changelog entry for this session, reconstruct my day, what did I ship. Rebuilds a work session from Littlebird capture into a dev log: repos and files touched, problems solved with what was tried and what worked, decisions that never became a commit message, a ready-to-paste Keep a Changelog block, and an honest coverage note. Session boundary is user-defined, not the calendar day. Reconciles against git when a repository is reachable.

Rebuild the sales pipeline that lives in your head. Trigger on 'what deals do I have open', 'rebuild my pipeline', 'my CRM is empty', 'who am I about to lose', 'which deals went cold', 'what did I quote and to whom', 'build my deal board', 'who hasn't replied to my proposal'. Reconstructs deals from captured DMs, discovery calls, proposals seen on screen, quotes sent and calendar holds into a stage-by-stage board with last-touch recency, a ranked going-cold list, and one next action per deal. Every stage is an inference and shows its evidence. Internal artifact only. Never contacts a prospect.

Guide a user through building their own personal writing-voice skill from their Facebook data export. Walks them through exporting their Facebook data (with screenshots), waits for the export, then unpacks, sanitizes, and analyzes their real posts and messages to produce a calibrated voice skill that writes exactly like them and reads 100% human. Use this skill whenever a user says "build my voice skill", "make Claude write like me", "clone my writing style", "create a voice skill from my Facebook", "analyze my posts", or uploads a Facebook data export zip. Not for brand voice guidelines for a company (use a brand-voice skill) - this is for one person's authentic personal voice.

Weekly attention forensics from Littlebird capture. Trigger on how fragmented was my week, where did my attention go, focus report, am I context switching too much, what broke up my week, rabbit hole check, weekly focus review, why did I get nothing done. Counts observed switches between work contexts, run length in consecutive snapshots, fragmentation by hour of day, named rabbit holes against stated intentions, and meeting load against unbroken calendar time. Ships week-over-week change plus one behavioral experiment to try. Never produces hours lost, percentages of the day, or a productivity score, because periodic snapshots cannot measure those. Self-analysis only, never pointed at another person.

Trigger on who has not paid me, chase unpaid invoices, accounts receivable, overdue invoice, aging report, payment follow up, outstanding invoices, invoice reminder, who owes me money, collections, past due. Rebuilds a receipted receivables table from Littlebird capture of invoices the user sent, reconciles it against every payment observation it can find, buckets what is left by age, forces a verification step before calling anything overdue, and drafts a polite-to-firm follow-up ladder per client that is held for approval and never sent.

Build my project knowledge base, document this project, write the PRD and architecture notes from my calls, make a docs pack for AI, what did we decide and when. Ingests one project's meetings, threads, on-screen artifacts and documents from Littlebird capture and produces a structured markdown documentation pack: product requirements, architecture notes, decision records, a glossary, a brand brief, and a register of every contradiction it found, each with both readings and both dates. Use for documenting a project that already exists in scattered conversations, not for writing a new spec from scratch.

Harvest hand-raisers from a comment-keyword campaign. Trigger on 'who commented my keyword', 'harvest my campaign leads', 'who DMd me about the launch', 'build my outreach list', 'comment to get it campaign', 'I missed a bunch of DMs', 'rank my inbound leads', 'who raised their hand'. Mines captured social activity across comments, DMs, friend and connection requests, and reactions, dedupes into one ranked roster with receipts and a coverage report, and drafts a first touch per segment for the user to send by hand. Drafts only. Never sends and never automates a platform action.

Stop re-debugging the same wall, capture what I learned, log this fix, add to my knowledge base, what did I figure out this week, I have solved this before. Watches Littlebird capture for moments the user got past something hard (an API error, a framework gotcha, a database quirk, a deploy failure) and files each one into a personal, greppable knowledge base of how they fixed it last time. Weekly routine that proposes candidate entries for confirmation, plus an on-demand pass after a hard session. Handles deduplication against existing entries, recurrence escalation, version staleness, and a mandatory secret scrub. Not for writing generic tutorials and not for documenting a procedure that worked first time.

Guide a user through building their personal writing-voice skill by mining their Littlebird data via the Littlebird MCP server. Covers installing Littlebird, connecting the MCP to Claude, mining the user's own verbatim writing and background context from captured screen history, chats, and meeting transcripts, confirming every retrieved fact with the user, and packaging a calibrated voice skill that writes exactly like them. Use this skill whenever a Littlebird user says "build my voice skill from Littlebird", "mine my Littlebird data", "make Claude write like me using Littlebird", or "what does Littlebird know about my writing style". Requires the Littlebird MCP (Pro plan). Not for company brand voice - this is one person's authentic personal voice.

Post-meeting follow-up. Trigger on "write the follow-up", "recap that call", "draft the follow-up email", "what did we decide", "send a recap to the client", "what did we not resolve", "sweep today's meetings". Turns one recorded Littlebird meeting into three things the built-in summary does not give you: a decisions record with the exact quote and timestamp behind each one, a sendable follow-up drafted in your voice and filtered for what those recipients should actually see, and a list of what got raised and never landed. Commitments are handed to the standing ledger. Runs on demand right after a call, or as a daily evening sweep. Requires the Littlebird MCP.

Audit SaaS and vendor spend from Littlebird screen capture. Trigger on subscription audit, where is my money going, cancel unused tools, find zombie subscriptions, failed payments, SaaS spend review, cut software costs, what am I paying for. Reconstructs a receipted vendor ledger from captured billing notices, receipts, dashboards and card alerts, proves which paid tools have not been opened in 90 days, traces failed-payment cascades back to one root cause, and produces a cancel, downgrade, consolidate and renegotiate action pack with cancellation emails drafted and held for approval.

Research a person, build a dossier on someone, vet a business partner, prep for a negotiation, verify what someone told you, or brief yourself before a meeting with a named individual. Fuses your entire Littlebird record of that person, every message thread, meeting, calendar invite, and screen sighting, into a dated relationship timeline with receipts, then reconciles it against external research. Use it for partner diligence, prospect research, negotiation prep, and claim verification. Not for employment screening.

Pre-call brief, meeting prep, prep me for my calls, what do I need to know before this meeting, who am I talking to, brief my day, call prep. Builds a one screen brief for every call on the calendar before it happens. Covers who is on it, what was said last time with the date, what each side committed to and whether it happened, objections they raised before, what changed on their side, three talking points and one thing not to forget. Runs as a nightly or early morning routine over tomorrow's calendar, or on demand for a single upcoming meeting. Requires the Littlebird MCP.

A 90-day radar for everything about to auto-charge. Trigger on what is renewing soon, upcoming renewals, cancel before it renews, domain expiring, SSL certificate expiry, annual subscription about to charge, auto-renewal deadline, renewal calendar, cancellation window. Mines Littlebird capture and calendar for renewal and expiry dates, builds a forward calendar of upcoming auto-charges with cancellation-window deadlines and evidence receipts, treats domains and certificates as their own class, and drafts cancel, downgrade and negotiation messages held for approval.

Research a topic and get the delta, not another wall of reading. Trigger on 'research this topic', 'what do I already know about', 'catch me up on', 'synthesize what I have read', 'is what I think about this still true', 'deep dive on a topic', 'what did I miss on', 'brief me on'. Pulls what the user has already encountered on screen and in messages about the topic, layers a fresh external sweep over it, and returns a synthesis split into what they already had, what is new since then, where the sources disagree, what they appear to believe based on what they actually said, and what is still open. Every external claim carries a URL. Optional recurring mode watches a standing topic and reports only what is new.

Audit my routines, why does my routine keep repeating itself, fix my routine prompt, my daily report is useless, set up a routine, what routines should I have. Scores every Littlebird routine against a nine-point rubric built from alert-fatigue research, diagnoses failures from the report history rather than the prompt text, writes complete replacement prompts with memory and escalation clauses, and wires each routine to the Cowork skill that resolves what it finds. Use for auditing, rewriting, or designing routines, not for running the work a routine reports on.

Mine meeting transcripts and captured writing for things you already said well, then draft posts from them. Trigger on 'what should I post', 'I have nothing to write about', 'build my content bank', 'turn my calls into content', 'find my best quotes', 'I said something good on a call last week'. Produces a weekly bank of 10 to 15 seeds, each carrying the verbatim quote, its receipt, the register it was said in, why it works, and a drafted piece, plus a mandatory do-not-publish list. Drafts through the user's own voice skill when one is installed. Never publishes and never posts.

What should I automate next, what skill should I build, am I doing the same thing over and over, I keep doing this by hand, find my repeated work, propose a new skill. Detects repeated manual workflows in Littlebird capture across five named signatures, ranks candidates with honest recurrence counts and receipts, dedupes them against the skills the user already has, recommends build or skip against ten reasons to say no, and pre-drafts a valid SKILL.md for the top candidate. Use for finding and drafting a new automation candidate. Not for auditing existing routines, and not for writing an SOP.

Write an SOP, document how I did that, turn last Thursday into a procedure, make this repeatable, write it up so I can hand it off. Reconstructs a standard operating procedure from Littlebird screen capture of work the user already performed, with the actual screens, field values, UI labels, and timestamps that were on screen, plus a named gap list and a mandatory redaction pass. Use for documenting a task that was already done, not for writing a generic tutorial.

Mines the praise you already earned into usable social proof. Trigger on "find my testimonials", "testimonial bank", "do I have any good client quotes", "social proof for the launch", "customer quotes", "can I use this review", "permission to quote a client", "case study quotes", "who has said nice things about us". Sweeps captured DMs, comments, reviews and meeting transcripts for praise, wins and results moments, then banks each as a verbatim quote with attribution, date said, a receipt and a permission tier: public and reusable, private and needs permission, or confidential and do not use. Separates results quotes into their own sub-bank with FTC typicality flags, drafts permission requests held for approval, and reports which client relationships have no captured praise at all. Internal artifact. Nothing is published and nothing is sent.

Weekly review, weekly scorecard, how was my week, week in review, weekly rollup, Sunday review, Friday review, what did I get done this week, next week's top three, weekly retrospective, weekly check-in. Composes the week's meetings and hours, commitments closed versus dropped versus open, leads captured, money findings, and content shipped into one scorecard by reading the sibling routines' own reports rather than re-deriving them. Leads with the multi-week trend rather than this week's figures, carries provenance on every number, is willing to state plainly that the week was poor, and selects next week's top three on consequence ahead of urgency with an escalate-or-drop rule for carried items. Runs as a weekly routine, or on demand. Requires the Littlebird MCP.

Finds the conversations you left hanging. Trigger on "who am I ghosting", "who am I leaving hanging", "what have I not replied to", "unanswered messages", "did I forget to reply to anyone", "cold threads", "who is waiting on me", "unreplied DMs", "weekly ghosting check". Sweeps Littlebird message capture and inbox snapshots for threads where you are the party who owes a response, ranks them on how directly the person addressed you, how much the relationship is worth, and what they were waiting for rather than on days elapsed, drafts a re-engagement line for each, and separates out the ones where the ball is actually in their court and the ones you should consciously write off.