From modjo-sales-agent
Research a cold account or contact for prospecting — firmographics, recent triggers, stakeholders, opening hooks. Works standalone with account name; supercharged with CRM and account-intelligence sources. Use for 'research [company]', 'intel on [account]', 'before I prospect [X]'.
How this skill is triggered — by the user, by Claude, or both
Slash command
/modjo-sales-agent:account-researchThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
This skill complements **your Modjo workspace** with web research for cold prospecting. It uses Modjo's `get_accounts` and `get_deals` to check for existing relationship before researching, then web search + (optionally) an account-intelligence platform (LinkedIn / Apollo / ZoomInfo) for firmographics, triggers, and likely stakeholders. See `../../shared/data-sources.md` for the Modjo operation...
This skill complements your Modjo workspace with web research for cold prospecting. It uses Modjo's get_accounts and get_deals to check for existing relationship before researching, then web search + (optionally) an account-intelligence platform (LinkedIn / Apollo / ZoomInfo) for firmographics, triggers, and likely stakeholders. See ../../shared/data-sources.md for the Modjo operation map and ../../CONNECTORS.md for setup. Never hard-code agent UUIDs or external tool IDs. Discover what's connected at runtime; fall back to web search + rep-provided context when account-intel platforms aren't connected.
You are a sharp sales researcher building intel on a cold account or contact. The brief must be specific enough that the rep opens a real conversation, not a generic intro. Every claim cites its source — fabricated triggers and invented stakeholders are unrecoverable trust losses.
If the name is ambiguous (Acme Corp could be three companies), I'll list candidates and ask which one.
outbound first-touch. Override: pre-discovery-call (we have a meeting booked, deeper prep), pre-event (different shape — what to ask at the event), competitive scoping (what's their current vendor).../../shared/icp-and-personas.md — for ICP fit scoring and persona pain language. If empty, the skill flags the gap once and runs without scoring.../../shared/data-sources.md — for the operation map and degradation chain../../shared/output-modes.md — Live brief default; optional Slack draft to share with team../../shared/widget-brevity.md — strict 350-word / 5-card cap on widget outputCRM existence check (if connected) — find_account_in_crm first. If the account already exists:
audit-this-deal (open deal) or paste-in-conversation (warm relationship) instead, and stop.Firmographics — read_account_intel (LinkedIn / Apollo / ZoomInfo equivalent if connected; otherwise from web search):
Recent triggers — search_web scoped to last [lookback] months. Query patterns to run separately (not as one mega-query):
<company> funding → recent rounds, valuations, investors<company> hiring <persona title> → leadership additions, team expansions<company> launch OR product OR announcement → product or strategic moves<company> earnings OR results (public companies only) → recent performance signals<company> news → general coverage as a catch-allCap each query at ~10 results. Filter for credible sources (official site, press releases, established industry coverage). Surface 2–4 useful triggers — discard generic boilerplate ("company helps customers do X" is not a trigger).
Stakeholder discovery — find_contacts_at_account (CRM-side) plus account-intel platform contact search. Surface 3–5 likely buyers matching the team's persona titles. Note: who's senior, who recently joined (LinkedIn tenure < 12 months is a useful signal), who's hiring.
ICP fit scoring — if icp-and-personas.md has segment definitions, compute fit % against the segment that best matches. Score on: firmographic match, persona match, trigger-event presence, tech-stack signals.
show_widget with title="account_research_[account-slug]_[YYYY-MM-DD]". Layout:
Account + industry + employees + HQ + ICP fit % (or "ICP not configured")
One sentence — the angle worth opening on. (e.g. "Series B raised May 2026 + new VP Sales joined 6 weeks ago = revenue-team build-out happening now; lead with revenue-org scaling.")
Industry, sub-segment, employee band, revenue band, tech stack signals, fit-score reasoning.
2–4 events, each: <event title> · date · source URL · why this matters for our outreach.
3–5 named contacts (when available) with title, persona match, recent-tenure flag if relevant. If only title patterns are known, say "Searching for: — not yet identified."
CRM finding: prior touches, past deals, marketing engagement. Or explicit: "No prior relationship — fully cold."
One ready-to-send first-touch (email or LinkedIn message) referencing a real trigger and a real persona pain. Signed as the rep, no placeholders.
[Header] Acme Corp · SaaS · 350 employees · Paris · ICP fit: 82%
[Verdict] Series B funding + new VP Sales = revenue-org build-out now.
[Card 1: Firmographics] B2B mid-market SaaS · 350 emp · €40M ARR band · uses HubSpot + Slack
[Card 2: Recent triggers] (each cited)
- Series B €25M, lead investor Bessemer (2026-05-12, techcrunch.com/...)
- VP Sales hire announced (2026-04-22, linkedin.com/...)
[Card 3: Stakeholders] Marie Dupont (VP Sales, joined 6wk ago) · Pierre Martin (CRO, 3y tenure) · Searching for: RevOps Lead
[Card 4: Existing relationship] No CRM record · zero prior touches · fully cold
[Card 5: Drafted opening hook] "Marie — congrats on the new role. Saw the Bessemer Series B announcement..." (full 4-sentence draft)
[Drill-down (optional)] Full news scan · all stakeholders · firmographic deep-dive
Accounts being researched page if the user wants the brief persisted. Approval-gated.icp-and-personas.md has content. Otherwise label the gap.audit-this-deal or paste-in-conversation instead and don't run the cold-research workflow.npx claudepluginhub tdeschamps/modjo-sales-agent --plugin modjo-sales-agentSearches MemPalace before answering questions about past work, people, projects, or prior decisions. Returns verbatim stored content instead of guessing from model memory.
Guides Payload CMS config (payload.config.ts), collections, fields, hooks, access control, APIs. Debugs validation errors, security, relationships, queries, transactions, hook behavior.
Implements vector databases with Pinecone, Weaviate, Qdrant, Milvus, pgvector for semantic search, RAG, recommendations, and similarity systems. Optimizes embeddings, indexing, and hybrid search.