djb258
- 34 skills
- 0 followers
- 9 hours ago last updated
- ▌ Barton Neon · djb258 bundleNeon Serverless PostgreSQL configuration, schemas, and operational patterns for barton-outreach-core — endpoint ep-ancient-waterfall-a42vy0du on us-east-1, database "Marketing DB", six production schemas (outreach, people, dol, company, cl, bit) plus vendor layer. Use this skill whenever querying, migrating, debugging, or making data architecture decisions in this repo. Trigger on: Neon, PostgreSQL, Postgres, database, schema, table, migration, connection string, pooler, outreach schema, people schema, dol schema, company schema, cl schema, bit schema, vendor_claude, enrichment pipeline, company-target hub, people-intelligence hub, dol-filings hub, blog-content hub, outreach-execution hub, coverage hub, talent-flow hub, or any reference to the relational data layer in barton-outreach-core. Also trigger on: column_registry, pipeline_errors, hub_registry, batch, or sub-hub architecture questions.
- ▌ Client Neon · djb258 bundleNeon PostgreSQL vault/archive configuration, schema topology, and operational patterns for the Client Intake & Vendor Export System — 5 spokes, 16 tables in the clnt schema, serverless driver over pooled connections, Doppler-managed secrets, Zod-validated intake pipeline, and vendor export egress. Neon serves as the vault/archive layer; CF D1/KV is the working database. Use this skill whenever querying, migrating, debugging, or making data-layer decisions about the Neon vault in the client repo. Trigger on: Neon, PostgreSQL, clnt schema, client_id, enrollment intake, vendor export, plan quote, election, invoice, spoke tables, column registry, codegen, promote-to-neon, or any reference to the vault database layer. Also trigger when discussing connection strings, migration scripts, error tables, staging tables, or the intake-to-canonical promotion pipeline.
- ▌ Client Lovable · djb258Figma UI design patterns, UI surface contract, and rendering constraints for the Client Intake & Vendor Export System — multi-tenant dashboard rendering via v_client_dashboard view, per-client branding, feature flags, and dashboard block layout. Use this skill whenever building, debugging, or making UI architecture decisions in the client repo. Trigger on: Figma, UI, dashboard, client portal, branding, feature flags, dashboard blocks, v_client_dashboard, multi-tenant rendering, domain mapping, or any reference to the presentation layer. Also trigger when discussing React components, Vite build, Radix UI, Tailwind, Shadcn, or the UI app structure. If the task involves what the user sees or how the client portal renders — this skill applies.
- ▌ Storage Neon · djb258 bundleNeon dev/secondary database for the Storage Container Go-NoGo feasibility analysis tool. Trigger on: Neon, dev database, secondary database, ZIP screening schema, zips_master, us_zip_codes, runs table, zip_results, stage_log, zoning_cache, api_cache, pricing_data, traffic_data, screening pipeline data, kill log, tier scoring, or any reference to the PostgreSQL layer that holds the ZIP code screening system. Also trigger when discussing database migrations, connection strings, branching for dev environments, or cold-start behavior on the storage screening backend. If someone mentions "the database" in this repo and it is about ZIP master data or screening run history, this skill applies.
- ▌ Storage Lovable · djb258Lovable-generated UI layer for the Storage Container Go-NoGo feasibility analysis tool. Trigger on: Lovable, lovable.dev, lovable-tagger, componentTagger, UI scaffolding, frontend generation, Vite config, ShadCN components, or any reference to the original app generation history. Also trigger when discussing the React frontend architecture, component library choices, or the export-to-GitHub workflow that produced this codebase. If someone asks "where did this frontend come from" or "why is the code structured this way," this skill has the answer.
- ▌ Neon · djb258 bundlePlatform capabilities, hard limits, pricing, and integration patterns for Neon Serverless PostgreSQL — connection pooling, branching, autoscaling, scale-to-zero, serverless driver, and edge function connectivity. Use this skill whenever building, querying, migrating, or making architecture decisions involving Neon or any PostgreSQL database in the stack. Trigger on: Neon, PostgreSQL, Postgres, database branch, connection pooling, PgBouncer, serverless driver, scale-to-zero, CU-hours, neonctl, or any reference to the relational data layer. Also trigger when discussing message queue tables, edge function database access, or source-of-truth data storage. If the task involves relational data that lives beyond D1's 10GB cap or needs full Postgres capabilities, this skill applies — even if the user doesn't mention Neon by name.
- ▌ Sales · djb258Master orchestrator for the sales intelligence engine. Routes /sales commands to the correct subagent or skill. Trigger: any message starting with "/sales". Flagship command: /sales prospect <url> launches 5 parallel subagents for comprehensive prospect analysis. All engines are industry-agnostic — the same orchestrator works for SaaS, insurance, real estate, or any vertical by loading the appropriate runtime config.
- ▌ Agent DB · djb258 bundleGarage Control Plane database governance agent — enforces database governance, produces DB_CHANGESET artifacts with migrations, rollback plans, and validation steps. Trigger on: WORK_PACKET with db_required=true, any mention of "db agent", "database migration", "schema change", "DB_CHANGESET", "drift detection", "column registry", "CTB validation", "migration plan", "rollback plan", "risk classification". This agent defines migrations but never applies them — the Builder applies. It validates registry-first compliance, cantonal cardinality, RAW immutability, vendor JSON containment, and bridge versioning before generating any changeset.
- ▌ Bootstrap · djb258 bundleStamps any repo with the IMO-Creator doctrine skeleton. Trigger: Developer runs /bootstrap in any repo root. Output: Doctrine-compliant file skeleton stamped for detected tier (Garage / Car / Sub-Hub). Non-destructive — fills missing files only, never overwrites existing doctrine.
- ▌ Sales Icp · djb258Ideal Customer Profile Builder — standalone skill invoked via `/sales icp <description>`. Takes a business/product description, conducts market research, and produces a comprehensive IDEAL-CUSTOMER-PROFILE.md with 6 ICP dimensions, negative ICP, 100-point scoring rubric, buyer personas, prospecting playbook, and competitive context. Industry-agnostic engine.
- ▌ Cloudflare · djb258 bundlePlatform capabilities, hard limits, pricing gates, and integration patterns for Cloudflare Workers, Hono, KV, D1, R2, Durable Objects, Queues, and Workflows. Use this skill whenever building, deploying, debugging, or making architecture decisions involving Cloudflare. Trigger on any mention of: Worker, Wrangler, D1, R2, KV, Durable Objects, Hono, Cloudflare, edge function, wrangler.toml, wrangler.jsonc, Hyperdrive, or any CF service name. Also trigger when the task involves the I/O boundary layer (Rim) in hub-spoke architecture — Cloudflare IS the Rim. If the code touches the outside world, consult this skill. Even partial mentions like "deploy to the edge" or "serverless function" should trigger this if Cloudflare is in the stack.
- ▌ Sales Prep · djb258Meeting Preparation Brief Engine — generates MEETING-PREP.md with 11 fixed sections covering company research, attendee intelligence, competitive context, and tactical preparation. Invoked standalone via `/sales prep <url>`. Industry-agnostic engine: section structure, question ordering, and response frameworks are constants; all prospect-specific content is runtime variable.
- ▌ Sales Report · djb258Sales Pipeline Report Generator — scans the working directory for prospect analysis files and compiles them into a unified, executive-ready pipeline report (SALES-REPORT.md). Synthesizes scores, classifies pipeline stages, computes health metrics, and produces prioritized action items. Industry-agnostic engine.
- ▌ Agent Auditor · djb258 bundleGarage Control Plane compliance agent — evaluates Builder execution against doctrine registry and audit rules across all lanes, issues classification, and feeds certification. Trigger on: execution ready for audit, any mention of "auditor", "audit execution", "evaluate compliance", "classify work", "certification gate", "check lanes", "fleet alignment check", "ORBT telemetry". Also trigger when processing inbox files from factory/runtime/inbox/auditor/. This agent is the FOURTH in the pipeline — it receives Builder output and determines PASS, FAIL_EXECUTION, or FAIL_SCOPE. It validates envelopes, checks lane artifacts, enforces fleet alignment, checks documentation staleness, emits ORBT error telemetry on failures, and verifies artifact integrity. Read-only — the Auditor never modifies files.
- ▌ Agent Builder · djb258 bundleGarage Control Plane execution agent — executes approved WORK_PACKETs across standard, DB, UI, container, and documentation lanes against mounted child repository clones. Trigger on: WORK_PACKET ready for execution, any mention of "builder", "execute work packet", "apply changes", "mount repo", "run lanes", "fleet refit", "apply refit bundle". Also trigger when processing inbox files from factory/runtime/inbox/builder/. This is the agent that DOES the work — it resolves repos, mounts clones, applies code changes within allowed_paths, applies DB_CHANGESETs, handles UI changes, runs containers, and produces documentation artifacts. Formerly named Worker (deprecated alias).
- ▌ Agent Planner · djb258 bundleGarage Control Plane planning agent — generates validated WORK_PACKET V2 envelopes and routes execution lanes. Trigger when: a Planner Intake Packet arrives from the Orchestrator, user intent needs to be translated into a WORK_PACKET, or any mention of "planner", "work packet", "route lanes", "generate WP", "plan execution", "scope the work". Also trigger when processing inbox files from factory/runtime/inbox/planner/. This agent converts structured intake into the execution envelope that drives the entire Builder pipeline. It validates target repos, classifies lanes (DB, UI, container, doc), and sets scope boundaries. If the work needs a WORK_PACKET, this is the agent.
- ▌ Sales Company · djb258Company Research Subagent — one of 5 parallel subagents for /sales prospect. Evaluates Company Fit (25% of Prospect Score) via firmographic analysis, technology detection, growth signals, and budget assessment. Industry-agnostic engine — works for any vertical via runtime config.
- ▌ Sales Qualify · djb258Lead Qualification Engine — evaluates a prospect against BANT and MEDDIC frameworks using public signals, producing an Opportunity Quality Score 0-100 with composite weighting (BANT 50%, MEDDIC 30%, Urgency 20%). Trigger via `/sales qualify <url>` standalone (outputs LEAD-QUALIFICATION.md) or as subagent of sales-prospect (returns score + structured data). Industry-agnostic engine.
- ▌ Skill Creator · djb258 bundleCreates validated, swap-testable skill packages for LLM execution. Trigger: User requests a new skill, or requests modification of an existing skill. Output: A validated skill package at skills/<skill-name>/SKILL.md that passes quick_validate.py and can be executed by any LLM without prior context.
- ▌ Lovable · djb258 bundlePlatform capabilities, hard constraints, credit economics, and pipeline position for Lovable.dev — the AI app builder. Use this skill whenever building with, exporting from, or making architecture decisions involving Lovable. Trigger on: Lovable, lovable.dev, AI app builder, vibe coding, rapid prototype, or any reference to generating full-stack apps from natural language prompts. Also trigger when evaluating whether a project should use Lovable vs traditional development, when dealing with GitHub export/import constraints, or when someone says "quick MVP" or "prototype fast" and Lovable is part of the conversation. If the question is "should we use Lovable for this?" — this skill has the answer.
- ▌ Sales Contacts · djb258Decision Maker Intelligence Engine -- identifies the buying committee, maps organizational hierarchy, researches personalization anchors, and builds a multi-threading engagement strategy. Trigger on `/sales contacts <url>`, or when launched as the sales-contacts subagent by sales-prospect. Standalone produces DECISION-MAKERS.md; subagent mode returns Contact Access Score 0-100. Industry-agnostic engine.
- ▌ Sales Followup · djb258Follow-Up Sequence Generator — produces multi-channel follow-up sequences for prospects after initial contact (meeting, demo, proposal, ghost, nurture). Trigger via `/sales followup <prospect>`. Industry-agnostic engine. Every email adds new value, references specific conversation points, and includes one clear next step. Output: FOLLOWUP-SEQUENCE.md.
- ▌ Sales Outreach · djb258Cold Outreach Sequence Generator — produces a complete, personalized, ready-to-send 5-email cold sequence with integrated LinkedIn touchpoints. Trigger via `/sales outreach <prospect>` standalone (outputs OUTREACH-SEQUENCE.md) or as the sales-strategy subagent within `/sales prospect` (returns Outreach Readiness Score 0-100). Industry-agnostic engine — outreach frameworks and sequence structure are universal constants; prospect details, triggers, and industries are runtime variables.
- ▌ Sales Proposal · djb258Sales Proposal Generator — produces a complete, client-ready CLIENT-PROPOSAL.md with 11 fixed sections, 3-tier pricing, 3 ROI scenarios, and a 6-email follow-up sequence. Invoked via `/sales proposal <client>`. Industry-agnostic engine. Deterministic template structure; LLM fills variable slots only.
- ▌ Sales Prospect · djb258Prospect Analysis Orchestrator — produces a scored, prioritized, actionable prospect report by launching 5 parallel subagents and synthesizing their results into a unified PROSPECT-ANALYSIS.md. Trigger on `/sales prospect <url>`, or when any task requires full multi-dimensional prospect scoring from publicly available data. Industry-agnostic engine.
- ▌ Sales Research · djb258Company Research & Firmographic Analysis Subagent — evaluates Company Fit (25% of Prospect Score) across 8 research dimensions and 5 scoring sub-dimensions using structured web intelligence. Trigger standalone via `/sales research <url>` producing COMPANY-RESEARCH.md, or as subagent 1 during `/sales prospect <url>` returning Company Fit Score 0-100. Industry-agnostic engine.
- ▌ Sales Strategy · djb258Outreach Strategy Subagent — one of 5 parallel subagents for /sales prospect. Evaluates Outreach Readiness (20% of Prospect Score) by synthesizing outputs from all 4 sibling subagents into an actionable outreach plan. Industry-agnostic engine — channel options and messaging frameworks are universal constants, prospect-specific decisions are runtime variables.
- ▌ Process Creator · djb258 bundleDocuments cross-repo business processes as PROCESS.md files. Trigger on: "document process", "write PROCESS.md", "process audit", "map the pipeline", "process definition", any request to create or update a process document. Also trigger when processes.ts flags a PROCESS.md as MISSING. Reads existing code, DB schemas, and pipeline definitions to produce a doctrine-compliant process document that maps the full IMO, ID chain, ERD, and migration status.
- ▌ Sales Objections · djb258Objection Handling Playbook Engine — generates OBJECTION-PLAYBOOK.md with word-for-word response scripts for 15 universal objections, 5 industry-specific objections, competitive battle cards, pricing tactics, and prevention techniques. Uses Feel-Felt-Found (FFR) and Acknowledge-Bridge-Close (ABC) dual-framework structure. Invoked standalone via `/sales objections <topic/industry>`. Industry-agnostic engine.
- ▌ Sales Report PDF · djb258Professional Sales Report PDF Generator — reads SALES-REPORT.md and prospect files from the current directory, assembles a structured JSON payload, and invokes a Python/reportlab script to produce a polished, multi-section PDF (cover page, score charts, comparison table, top prospects detail, action plan, methodology). Generic engine: PDF layout, sections, color scheme, and JSON schema are constants. Company names, scores, and prospect data are variables.
- ▌ Sales Competitive · djb258Competitive Positioning Subagent. Evaluates Competitive Position (15% of Prospect Score). Detects incumbent tools, assesses switching costs, identifies feature gaps, builds positioning angles, produces a battle card. Industry-agnostic — no hardcoded verticals.
- ▌ Sales Competitors · djb258Competitive Intelligence Engine Subagent — detects what tools/solutions a prospect currently uses via 6 detection methods, builds 10-section battle cards per overlapping competitor, and assembles a scored competitive strategy. Trigger standalone via `/sales competitors <url>` producing COMPETITIVE-INTEL.md, or as subagent during `/sales prospect <url>` returning Competitive Position Score 0-100. Industry-agnostic engine.
- ▌ Sales Opportunity · djb258Opportunity Assessment Subagent — evaluates Opportunity Quality (20% of Prospect Score) using the BANT qualification framework against public signals. Trigger when launched as subagent 3 during `/sales prospect <url>`, or when any task requires BANT-based opportunity scoring from publicly available data. Industry-agnostic engine.
- ▌ Agent Orchestrator · djb258 bundleGarage Control Plane intake agent — deterministic normalization, ID minting, ORBT classification, and routing to Planner. Trigger on any new work request entering the Garage pipeline: user intent that needs to become a WORK_PACKET, raw task descriptions needing structured intake, or any mention of "orchestrator", "intake", "mint ID", "classify ORBT", or "route to planner". Also trigger when processing inbox files from factory/runtime/inbox/orchestrator/. This agent is the FIRST in the pipeline — nothing reaches the Planner without passing through the Orchestrator.