modootoday

Your own Naver blog, Place, Smart Store, shopping, search ad and news comment data read through the datalab.tools browser extension, plus card news, short-form video, tutorial posts and social copy built in its editors, with no invented numbers and no change made without asking.

31 skills

Numbers that mean what they say: ClickHouse SQL that reads the key and deduplicates correctly, Naver trend data read as ratios, metrics defined once, experiments designed before launch, retention read by cohort, and open survey responses grouped with every response id kept.

16 skills

Reading passages that hit a measured grade level, language drills, and teacher-side checks for coverage, answer keys, lesson timing, study calendars and rule-valid puzzles, with the numbers and logs shown.

14 skills

Growth and community work: store listings written for the people searching that store, Discord servers run with structure and consistent moderation, newsletters readers open and finish, a month of social posts a small team can ship, launch emails that lead with what changed, and sites submitted to Google, Naver and IndexNow engines from a pipeline.

13 skills

Early product-planning steps done carefully: spar on an idea, frame the problem behind a request, synthesize feedback without overstating it, rank a backlog while saying where the ranking is fragile, write a Korean PRD with testable acceptance criteria, and put a decision on one page with the recommendation first.

13 skills

Korean content that reads well and stays true: product copy fixed to the Toss writing principles, Naver blog posts planned for search, persona tone measured with markers, and ghostwriting that never invents the author's facts.

11 skills

Video work that can be executed and verified: briefs and storyboards someone else can produce from, Remotion explainers rendered from frame math, talking-head edits with ffmpeg that keep a list of every cut, and Korean subtitles timed to the voice by forced alignment.

11 skills

Images that hold up at their real size: share images rendered from HTML with Korean text that wraps, consistent hand-written SVG icon sets, batch post-processing that crops, compresses and strips location data, and product shots edited from the real product photo with a ledger and a label check.

10 skills

Reviews of platform pieces that fail in production for configuration reasons: Cloudflare Worker bindings, Manifest V3 extension lifecycles, Chrome Web Store policy before submission, MCP tool design and Postgres migration locks.

10 skills

Keep test cases aligned with change: a trace from requirement to changed behavior to case, with gaps and duplicates listed and automation proposed for QA approval.

9 skills

Check AI-suggested citations, measurement instruments and extracted statistics against real sources before they enter your research notes, and redo lab dilution and unit arithmetic before results are used.

8 skills

Audio from brief to file: a brand's sonic directions, scripts written for the ear, Korean TTS with spoken numbers, mixing to a measured loudness, and consistent sound effect sets.

7 skills

Engineering document procedures: work instructions from field notes, clause location with verbatim quotes, sourced spec tables and fault evidence logs.

7 skills

Quality checks for translations and subtitles: omissions, footnote parity, naturalness, and cue limits reported by id.

7 skills

Serve the skills your installed packages already ship, and the skills in this marketplace: read SKILL.md out of node_modules and plugin marketplaces, and answer with the one skill a question needs instead of the whole catalog.

6 skills · 1 MCP server

Checks for books before release: typeset text diffed against the approved manuscript, print and EPUB rules kept apart, automated accessibility results paired with the manual items they cannot prove, a project style sheet applied across chapters, index locators checked against page proofs, and new chapters of long fiction checked against the author's canon ledger and outline with every ruling left to the author.

7 skills

Write agent skills that change behaviour: evals before the body, descriptions that trigger on the right requests, rules tied to the failures they prevent, and a measured delta before release.

7 skills

Write plans, source-of-truth documents and decision records that can be proven stale, superseded without deletion, and told apart from one another.

4 skills · 3 commands

Coordinating subagents: whether and how widely to delegate, briefs with boundaries and hand-back schemas, partitioned writes with an idempotent merge, verifying what a subagent claims, fair comparisons, and fan-out scripts with caps and resume.

6 skills

Brand work that decides something: visual directions argued before drawing, a tokens.json and colour names every asset reads, client requests clarified before design, mockup text recovered with its doubts marked, and QA of the rendered asset before it is published.

6 skills

Prove that migrated, converted or rewritten SQL returns the same result as the original: statements mapped one to one, two-way key diffs, and checkers shown to catch seeded differences.

6 skills

Day plans, seating plans, deadline registers, hotel shortlists and route checks that survive arithmetic: fixed bookings, opening hours, travel time, supplied rules and official sources, with buffers, backups and unverified items marked.

6 skills

Checks and plans for GIS data work: raster NoData and integer scaling, geometry validity and merges, label expressions and spatial views, slope areas from contours, open replacements for licensed tools, with CRS and units stated.

6 skills

Hobby help that keeps its facts straight: game state and rules kept legal across chess, board game and family role-playing sessions, and crochet or knit patterns with every stitch count recomputed and every part joined.

6 skills

Household paperwork and planning on the facts a person supplies: pay-period schedules where every dollar is placed once, receipts split to the total, a small-step week for chores, a ledger of home service events, total-cost comparisons of personal offers, a trading journal checked against your own rules, and insurance bills and payments reconciled per claim.

6 skills

Customer operations: Korean support tickets triaged with incidents spotted across them, status updates that say what is affected and when the next update comes, read-only runbooks for on-call, and blameless postmortems with a sourced timeline and owned actions.

6 skills

Procurement working procedures: spend projection, dedicated lane cost models, freight classification lookups, shortlist verification, request intake with stage checks, and ERP navigation with verification notes.

6 skills

Habits for long agent sessions: status briefings with proof, follow-up work run in chosen waves with standing rules kept, subagent fan-out from one brief, and one living document per workstream.

5 skills

Decks that argue something and survive review: a storyline built from the answer down, pptx files filled from the template's own layouts with native charts, a QA pass for numbers that disagree, question rehearsal for the room, and edits that keep content and hand fixes.

5 skills

Procedures for working across agent runtimes: check remaining allowance, hand off bounded jobs through files, research with cited peer reports, compare subscription and token costs, and verify an adversarial second opinion.

5 skills

Make an existing pile of design documents navigable: check their schema, naming and placement, group the ones sharing a subject, draw the graph they form, and guide review of stale ones.

5 skills

Read images, scans, charts and device screens for blind and low-vision users: the text first, in order, with trends stated and uncertainty marked.

4 skills

Keep one rule source per package and project it into the filename each coding agent expects, with a lint for rule documents that have started holding something else.

4 skills

Refactors of analysis pipelines that must keep results equal: baseline frozen, steps mapped, outputs compared at a stated tolerance, differences explained apart from unexplained ones.

4 skills

Seller checks: second-hand items identified from photos and markings, listing copy kept to each SKU's spec row, reorder points from a sales sheet, and shopping-feed disapprovals triaged without invented identifiers or prices.

4 skills

Construction checks: defect and warranty lists split by trade, code questions answered only from pasted code text, takeoff quantities and units recomputed, and field questions turned into single-issue RFIs with a log.

4 skills

Design handoff checks: every wireframe element mapped to a named component of the design system before a prototype is built, built UI diffed against the design's values, colour contrast of component states computed, and handoff state matrices whose missing states become questions for the designer. It plans and checks only; it builds nothing.

4 skills

Workout log tables and equipment substitutions by movement pattern, with every load change left for a trainer to approve.

4 skills

Review MusicXML imports and proposed edits against explicit playback paths, sounding-pitch references, percussion bindings and approved beat boundaries while preserving musical content.

4 skills

Abuse detection that can be trusted: signals ranked by how abusers work and how cheaply they could evade, rules measured on labelled data before they act, and appeals that restore what a false positive took.

4 skills

Blender scenes planned and checked against numbers: dimensions, connections and collisions per state, reference comparison, and web delivery budgets.

3 skills

Run Claude in Chrome work from any session: directly when it has the browser, otherwise by a request to a session that does, with answers in one fixed format.

3 skills

Build benchmark controls, portability reviews and toolchain equivalence pilots.

3 skills

Entertainment-use readings with the arithmetic kept honest: numerology, chart and spread inputs computed only from the rules you name, every step shown, and no predictions.

3 skills

An assistant that stays inside the help you chose when you talk through a hard time: it records the scope, asks before changing mode, reads back only your own step list and never diagnoses or reframes.

3 skills

Supplied financial figures checked by arithmetic: statement rows tied out to opening and closing balances, documents inventoried with evidence, and cash flow movements bridged by driver.

3 skills

Home and kitchen recipe work that keeps the source intact: a pasted recipe copied to a card without invented changes, then every ingredient swap or appliance change logged as a marked, estimated adaptation.

3 skills

Family-history research that marks what it does not know. It indexes a surname across volumes with page context, flags same-name risk and lists unreadable pages as gaps, offers uncertain readings of damaged handwriting with the comparison letters used, checks tree exports for impossible dates and relationships without editing them, and keeps evidence, interpretation and open questions apart. Work is on pasted records only; use synthetic or historical data, not living people.

3 skills

Measure which lifecycle hook events actually fire in each agent runtime, and prove context injection and stop-blocking really reach the model.

2 skills · hooks

Browser simulations and visual demos checked against a known result: error against an analytic or reference value, timestep convergence, unit consistency between views, and frame time reported only for the devices actually measured. It verifies a demo; it does not build one.

3 skills

Let two already-running coding-agent sessions talk to each other at turn boundaries, with no daemon and no proxy, over a mailbox both sides agreed to open.

2 skills · hooks

Photo cull and batch-grade checks from the scores, logs and settings you supply: subject-specific criteria, rank-and-flag only, preview-versus-export match and file-to-output mapping. It checks and plans; it does not edit or generate images.

3 skills

Media targets and pitch angles for a PR pitch, checked against what was actually pasted: journalist beats and bylines marked verified or unverified, no contact details that are not on a source page, calendar deadlines compared with today and hooks tied to approved expertise only. It prepares a checkable worksheet; it does not write the pitch or profile anyone.

3 skills

Screening and records for property work: listings compared with a buyer's must-haves with unknowns kept unknown, room layouts checked against fixed structure, and a tenant's deposit or move-out dispute arranged as a dated, referenced evidence packet with no legal conclusions.

3 skills

Sales follow-through from the record: forgotten promises recovered from sent email with dates resolved and the quoted line, and CRM framework fields and a follow-up email filled only from call transcript evidence.

3 skills

Carousel posts and card news from copy to files: slide-by-slide copy with a hook and one call to action, slides rendered from HTML at platform sizes, and Korean card news for Naver with the post text around it.

3 skills

Index interview transcripts by session, theme and timestamp with verbatim audited quotes, score SUS questionnaires with the scoring rule the user supplies, and draw an information-architecture sitemap from a feature list.

3 skills

Refuses a small, measured set of shell commands whose damage is hard to undo,

1 skill · hooks

Cap the concurrency of build and test commands an agent starts, advising by default and rewriting only when asked.

1 skill · hooks

Checks fine-tuning data against the target trainer or provider schema before an upload or a training run: roles, field names and content types per row, one JSON record per example, a reload with the real loader, and tool-call turn structure. It checks data; it does not train.

2 skills

After an edit, runs the check belonging to the package that was edited, once per

1 skill · hooks

Refuses shell commands that would print a credential file or send one over the

1 skill · hooks

Review archive extraction boundaries and applied PDF residual-content evidence with exact artifact, API and inspection limits.

2 skills

Tell a session which checkout it is standing in, and whether other working trees or another session's uncommitted files are present.

1 skill · hooks

Permission setups for agents that run without a person watching: approval branches walked, denied actions held across rewordings, credentials kept out of the model and success checked on the artifact.

1 skill

Decide what belongs in the system prompt of a tool-using agent and what to leave out, with the symptom each omission produces.

1 skill

Synthetic isolation fixture

1 skill

Fact-preserving drafting of clinical notes: a clinician's own brief notes placed into a required template, empty fields marked NOT DOCUMENTED, model wording listed for the clinician to confirm.

1 skill

Sourcing search strings built from stated job requirements, with protected-trait terms and their proxies left out. Nothing here ranks, scores, screens or profiles individual people.

1 skill

A fix made on an anonymized copy of an administration script patched back to the real script through a local replacement table, changed hunks only, with no real value sent to a model.

1 skill

AI-generated content and ads labelled per platform and current Korean rules, with a record of how each asset was made.

1 skill

Draft billing time entries from a day's work notes that follow the client billing guidelines the user pastes, keeping the stated total, citing the rule behind each change and listing entries the guidelines would reject.

1 skill

Prices framed as hypotheses with a test: who pays, for what unit of value, against which alternative, what result would change the decision, and the cheapest test before anyone builds a pricing page.

1 skill

Report which of a project's written agent rules are actually enforced by a hook and which hold only while someone remembers them.

1 skill