hectorai-live
OrganizationHector's Amazon Ads methodology as Claude skills — brand ads audit, keyword harvesting, SQP analysis, and search-term negation, running on the Hector Amazon MCP connector.
Categories
Indexed Skills (5)
hector-amazon-ads
Router and shared operating rules for the Hector Amazon Ads skill set. Dispatches to four SME-authored skills — brand ads audit, keyword harvesting, SQP analysis, and search-term negation — and holds the MCP conventions they share (profile resolution, date handling, row caps, filter placement, units, write gates). Load when more than one Hector Amazon skill could apply, or when a request touches Amazon Ads via the Hector MCP but does not match a specific skill's trigger.
hector-brand-ads-audit
Use ONLY when the user explicitly asks to run a "brand ads audit", "ads audit", "advertising audit", or "brand advertising audit" for a named brand or Amazon Ads profile. Trigger phrases include "run an ads audit for X", "run a brand audit for X", "do an advertising audit on X", "audit X's Amazon ads", and close paraphrases where "audit" + ("ads" OR "advertising" OR "brand") co-occur. Builds a 5-pillar HTML slide deck (Account Overview, Bleeding, Harvesting, Headroom, Placement + Strategy) via Hector MCP and renders 1280×720 slides. Does NOT fire on conversational mentions of audits, on bid-lookup or date-resolution requests, on questions about audit methodology without an account named, or on requests to edit this skill. ALWAYS asks precursor questions (account+country, date range, comparison toggle, focus ASINs, currency-override) before any MCP call.
hector-keyword-harvesting
Use this skill ANY TIME the user asks about keyword harvesting, finding keywords to add, identifying search terms not in exact match, or gap analysis — e.g. "harvest keywords", "what keywords should I add?", "which search terms should I be targeting in exact match?", "find terms with orders not in my exact targeting", "run a harvest for [brand/account]", "show me harvesting opportunities", "keyword gap analysis", or close paraphrases. Also fires when the user wants to add the harvested terms to existing or new campaigns (action phase). Does NOT fire for: "show me my top search terms" (performance report only), "which search terms are wasting spend" (negative keyword task), "what keywords should I pause/remove" (bid or status task), general search-term analytics with no intent to promote terms, placement analysis, or budget questions.
hector-search-term-negation
Use ONLY when the user explicitly asks to add negative search terms based on bleeding data — trigger phrases include "run search term negation for X", "negate bleeding search terms for X", "add negatives for bleeding in X", "execute bleeding negation on X", and close paraphrases where "negation" or "negative" + a named brand co-occur. Identifies bleeding search terms per the standard Hector bleeding rules (SP-only, ≥ floor(1/CR) clicks, 0 orders, exclude branded, exclude ASIN-pattern, ungrouped), then negates each identified SearchTerm as a NEGATIVE EXACT MATCH in the specific ad group where it is bleeding. Does NOT fire on conversational mentions of negation, on single-term negation requests, on audit requests, or on requests to edit this skill. ALWAYS presents a preview and requires explicit CONFIRM before any write. Runs independently of the audit skill — no prior audit required.
hector-sqp-analysis
Use ONLY when the user explicitly asks to run an "SQP analysis", "search query performance analysis", "SQP report", or "search query report" for a named brand or Amazon Ads profile. Trigger phrases include "run SQP analysis for X", "do an SQP report for X", "search query performance for X", "analyse SQP for X", and close paraphrases where "SQP" OR "search query performance" co-occur with a brand name or account. Builds a 6-tab HTML dashboard (Tab 0 = cover/summary, Tabs 1–5 = analysis) and an Excel workbook (5 data sheets) covering: Movers & Shakers (WoW search-volume trend), Top Keywords (brand funnel snapshot), Scale-Up Keywords (dynamic IS% threshold based on brand average), AOV Gap Analysis (brand AOV vs category AOV gap ≥15%), and Funnel Drop-Off (brand funnel stage classification). Token-efficient by design: pulls ONLY the 10 metrics the outputs use, applies server-side metric_filters to drop long-tail terms before they enter context, reads file-offloaded tool results from disk instead of re-transcribin
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.