understand-anything
claudeAI-powered codebase understanding — analyze, visualize, and explain any project
9 skills · 10 agents · hooks
MixShift AI skills for Amazon advertising and retail operations. Query the MixShift warehouse (Sponsored Ads, DSP, Seller and Vendor Central revenue, orders, inventory, catalog) with read-only SQL and CSV export. Pull live Amazon data: SP-API reports, retail lookups, AMC and DSP analytics. Run PPC optimization workflows: health checks, bid management, search term negation, monthly reports, brand setup. Ads writes preview first and require explicit confirmation.
Plugin mixshift-ai from marketplace mixshift in miXshift/mx-claude-plugin. Commands read from the repo's manifest — 2 days ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
mx-amazon-adsRead and write the live Amazon Ads API for any advertiser the signed-in MixShift tenant is connected to, straight from Amazon, through the bundled harness CLI. The Ads analogue of mx-data-explore and mx-amazon-report: a catalog-driven workhorse for the current state of an ad account (campaign, ad gr
mx-amazon-amcRun ad-hoc Amazon Marketing Cloud (AMC) clean-room SQL analytics for a merchant, straight from Amazon through MixShift's service. Owns the full AMC loop: find the AMC instances a login can reach, inspect the data sources, help write dialect-correct AMC SQL, submit a workflow, poll it across turns, a
mx-amazon-dspRead and build an Amazon DSP advertiser straight from Amazon. REPORTING: how DSP did last year, history older than 90 days, year-over-year or monthly history. ACCOUNT READS: what is set up now: campaigns, line items, creatives, which creatives sit on which line items and whether they are live, appro
mx-amazon-inboundWalk an FBA inbound shipment from start to finish, one confirmed step at a time: create the plan, packing, placement, carrier, labels and tracking, or cancel a plan. Every fee Amazon would charge (placement, a partnered carrier, a late cancel) is shown to you first and committed only after you say y
mx-amazon-reportPull Amazon SP-API reports on demand, directly from Amazon, for any merchant and window you need. This is the report analogue of mx-data-explore: a catalog-driven workhorse for fetching data MixShift may not already hold in the warehouse, or a known report for a specific ad-hoc time frame, so the us
mx-amazon-retailLive Amazon SP-API retail lookups, straight from Amazon, for any merchant you are authorized for. Single fast request/response calls (not report documents; that is mx-amazon-report) for point-in-time questions about your catalog, inventory, orders, finances, listings, and fulfillment. The headline u
mx-amazon-shelfSee the Amazon shelf the way a shopper at a given zip sees it, on demand: who holds the Buy Box for an ASIN and with what delivery promise, every offer on a listing with each seller's price, where ASINs rank for a keyword (organic and sponsored, desktop and mobile) with badges, coupons and stock war
mx-asin-target-negationPhase 2 negation review for ASIN targets matched through auto campaigns, category targeting, and other Product Attribute Targeting paths. Pulls ASIN-triggered rows for a configurable window, suppresses ASINs already under manual targeting, evaluates PDP/form-factor overlap, joins lifetime performanc
mx-auth-loginAuthenticate against the MixShift warehouse via the mx-legacy-auth service at mcp.mixshift.io. Routes by scenario: interactive sign-in (device-code or PKCE browser flow) when a human is present, service credentials (admin-issued, OAuth client_credentials) for unattended runs like scheduled Cowork ta
mx-auth-service-setupSet up an admin-issued service credential so MixShift skills can run UNATTENDED: scheduled Cowork tasks, daily crons hitting the warehouse or the SP-API report/pricing endpoints, CI jobs, cloud automations. The default flow needs no secret handling at all: the admin mints a one-time SETUP CODE at Pe
mx-brand-contextSet up Brand Context for a brand: build its Tier 3 brand context for a new Amazon PPC account from scratch (a typed context.yaml, narrative.md, corpora/, and a reviewable Brand Context page). Use it whenever the user mentions setting up or building brand context, onboarding a brand, a cold start, ne
mx-budget-balanceThis skill should be used when the user asks to "run a budget balance", "balance my budgets", "weekly budget check", "which campaigns are capped", "which campaigns need more budget", "where should budget move", or wants the weekly Sponsored Products budget-mix pass, or says "demo budget balance" for
mx-daily-health-checkThis skill should be used when the user asks to "run daily health check", "check account status", "how is the account", or needs comprehensive daily exception-based analysis of advertising account performance across campaign types, objectives, and item groups. Uses percentile-based confidence interv
mx-data-exploreAd-hoc query, sample, and CSV export against MixShift's MySQL warehouse (legacy schema: dashamazon). Covers Amazon Sponsored Ads (SP/SB/SD), DSP, Seller Central + Vendor Central operational revenue and orders, inventory, and dimensional catalog tables — not just PPC. Read-only, routes through the bu
mx-featured-offer-checkCheck, on demand, which of your ASINs hold the Amazon featured offer (the Buy Box) and which do not: Amazon's own answer per shopper segment with its share of views, then the shelf for every ASIN you lose, so each loss is explained (a faster delivery promise, a lower price, no featured offer at all,
mx-feedbackUse this when the user wants to send feedback to MixShift, or when they express a complaint, frustration, or dissatisfaction about the plugin even without using the word "feedback". Examples: "this is broken", "this is way too slow", "it feels broken", "this isn't working", "this is confusing", "I w
mx-helpNavigation and discovery hub for the MixShift plugin. Use when the user asks for help, says they are stuck or lost, asks what the plugin or a specific skill can do, asks how to do something or where to find something, wants to browse or list the available skills, or wants to know how to fix a proble
mx-keyword-bid-healthThis skill should be used when the user asks to "run keyword bid review", "bid health", "weekly bid check", "bid optimization", or needs weekly keyword-level bid optimization review. Surfaces high-ACOS bid reduction candidates and scale opportunities with proven conversion volume. version: "0.5.1" a
mx-max-forecastThe max tier of forecasting: a revenue forecast for one Amazon account, or for one brand inside an account (a sub-brand), explained in plain language with its measured accuracy, from the MixShift Intelligence service. Serves the forecast the MixShift forecasting app published (corrections included)
mx-max-reportThe max tier of MixShift reporting: prepares a client-ready performance brief and a private internal companion for any account, on any cadence (monthly, bi-weekly, QBR). Core figures come from the MixShift Intelligence service; the warehouse battery adds the daily series, availability, page-view-wei
mx-monthly-reportGenerates and publishes a monthly Amazon advertising performance report for a brand in MixShift's canonical analytical voice. Covers MoM and YoY comparisons, H-Bridge efficiency analysis, item group highlights, forecast beat/miss, and a Looking Ahead section. Saves report locally as /monthly-report.
mx-phrase-negative-discoveryDiscovers phrase negative candidates by decomposing search term corpus into n-grams, aggregating performance per sequence, and surfacing sequences with combined spend and zero conversion history. Applies conflict detection, semantic clustering, and brand context filtering. Phrase negatives have blas
mx-portfolio-quick-scanMulti-account daily triage. Produces one status card per account answering a single question: do I need to log into the platform today? GREEN / YELLOW / RED verdict per account based on CI anomalies, TACOS posture, and pacing proximity. No dimensional tables. No keyword data. Replaces on each run —
mx-ppc-relevance-checkSemantic relevance classification for search terms and ASIN targets. Operates on a provided list of search terms or the Tier B/C set from phase 1 exact-term negation. Separate from threshold logic — a term can be low-spend AND relevant (don't negate) or high-spend AND irrelevant (negate for differen
mx-runaway-spend-checkThis skill should be used when the user asks to "run runaway spend check", "check for runaway keywords", "identify overspending keywords", or needs daily keyword-level acute runaway spend detection. Flags keywords where T-1 spend spiked materially or where high-spend keywords generated zero conversi
mx-scheduled-taskBuild or repair a scheduled MixShift task that keeps working with nobody at the keyboard: a Cowork scheduled task or a Claude Code cron that pulls warehouse data, Amazon reports, health checks, or portfolio scans on a schedule. Owns the whole lifecycle: pick the durable anchor folder the task can al
mx-search-term-data-pullPure data extraction layer for search term analysis. Pulls window ST report, lifetime performance at three granularity levels, existing negatives, and manual ASIN targets. Outputs structured JSON artifact for downstream skills. No judgment, no LLM analysis — data only.
mx-search-term-harvestIdentify search terms and ASIN targets converting efficiently under auto or broad match that are not yet in explicit keyword targeting. Promotes high-performing auto terms to explicit campaigns. Pure harvest logic — no negation, no relevance judgment.
mx-search-term-negationUniversal workflow for Amazon PPC search term irrelevance analysis and negative keyword strategy. Use when conducting systematic review of search term performance to identify irrelevant traffic and implement surgical negative keywords.
mx-share-skillUse when the user wants to share or submit a skill they built with or for the MixShift plugin, contribute a skill to the library, or send MixShift a skill, workflow, or reusable prompt they made. Also use when the user changed one of the plugin's own skills and wants that change considered. Bundles
mx-updateUse when the user asks to update the plugin, update MixShift, says they are on an old version, just updated, or wants to catch up after an update or see what to do with new features. Shows the exact update steps for their surface (terminal, Claude Code, or Cowork), what changed via `mixshift whatsne
mx-walmart-adsRead and manage Walmart ads (Walmart Connect Sponsored Search, plus Display where the account has it) for the Walmart advertisers connected to the signed-in MixShift account, straight from Walmart, through the bundled harness CLI. The Walmart analogue of mx-amazon-ads: the advertisers you can use, c
mx-walmart-retailBeta. Read your Walmart Marketplace seller account for the Walmart sellers connected to the signed-in MixShift account, straight from Walmart, through the bundled harness CLI. Gives one view per SKU of what is selling, what you have, what is inbound, how many days of supply you have left and what to
mx-welcomeShow the MixShift plugin first-run welcome message. Use when the user says "welcome", asks "what is this plugin", asks how to get started, or is new and needs orientation on credential retrieval + initial setup steps. version: "0.1.2" author: "MixShift" - welcome - mixshift welcome - run welcome - w
node "${CLAUDE_PLUGIN_ROOT}/hooks/session-start.mjs"
AI-powered codebase understanding — analyze, visualize, and explain any project
9 skills · 10 agents · hooks
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
Rewrite AI-sounding text so it reads like the writer without changing what it says.
1 skill
AI memory system — mine projects and conversations into a searchable palace. 47 MCP tools, auto-save hooks, guided setup.
3 skills · 6 commands · 1 MCP server · hooks
Perform AI code review on Git diffs — supports workspace changes, branch ranges, and single commits with concurrent per-file analysis, codebase search, and deep context-aware review.
2 commands
Eval-gated LLM fine-tuning lifecycle (LoRA/QLoRA, DPO, GRPO/RLVR, vision SFT, quantized export). Use when fine-tuning an open-weights model end-to-end: building the eval harness first, selecting method and model, preparing data, training with Unsloth or TRL, gating checkpoints, and exporting GGUF/FP8.
10 skills · 2 commands · 3 agents
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.