--- name: Cold Email Engagement-First Outreach description: Root skill for planning, drafting, auditing, and routing cold outreach to strangers across B2B sales, founder-led outreach, investor fundraising, recruiting, partnerships, PR, podcasts, and customer research. Use when the user needs a campaign bundle, one targeted email, reply handling, or a trust-preserving outreach operating workflow. legacy_paths: - cold_email_outreach - cold-email-engagement-first-outreach - sales-and-growth/cold-email-engagement-first-outreach - growth.cold_email.skill - docs/research/youtube-library/skill-drafts/cold-email-engagement-first-outreach/SKILL.md child_skills: - id: cold_email_icp_signal_design name: Cold Email ICP and Signal Design summary: Deep child skill for defining the right person, right moment, segment, buying signal, and disqualifiers before outreach. when_to_load: - When the target list, ICP, persona, or timing signal is vague, mixed, or untested. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_icp_signal_design/SKILL.md - id: cold_email_offer_lab name: Cold Email OfferLab summary: Deep child skill for creating or repairing the artifact offer and smallest useful yes. when_to_load: - When the user has no offer, has a meeting-first offer, or needs to design a free useful artifact. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_offer_lab/SKILL.md - id: cold_email_research_anchors name: Cold Email Research Anchors summary: Deep child skill for mining one specific, real, recent anchor and bridge for strategic or single-target outreach. when_to_load: - When strategic, single-target, investor, recruiting, PR, podcast, or partnership outreach needs stronger research. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_research_anchors/SKILL.md - id: cold_email_outreach_compiler name: Cold Email Outreach Compiler summary: Deep child skill for compiling prepared mode, segment, anchor, offer, proof, and sender constraints into a finished outreach bundle. when_to_load: - When the inputs are ready and the user wants a finished email, campaign bundle, or audit rewrite. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_outreach_compiler/SKILL.md - id: cold_email_taste_review name: Cold Email Taste Review summary: Deep child skill for reputation-risk review, bridge quality, proof integrity, and whether a serious person should send the draft. when_to_load: - When a draft may be generic, pushy, over-automated, weakly bridged, or reputation-risky. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_taste_review/SKILL.md - id: cold_email_deliverability_readiness name: Cold Email Deliverability Readiness summary: Deep child skill for sender trust, cold-domain readiness, volume caps, warmup, bounce risk, and complaint safeguards. when_to_load: - When sending at scale, using cold domains, or diagnosing low opens, bounces, spam placement, or complaint risk. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_deliverability_readiness/SKILL.md - id: cold_email_reply_os name: Cold Email Reply OS summary: Deep child skill for classifying replies, routing objections, reviving silence, and preserving trust after response. when_to_load: - When a reply arrives, a thread goes quiet, objections are expected, or the user needs reply-to-call handling. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_reply_os/SKILL.md - id: cold_email_learning_review name: Cold Email Learning Review summary: Deep child skill for turning campaign metrics, objections, and buyer language into next tests. when_to_load: - When a campaign or test has results and the user needs to decide stop, iterate, recycle, or scale. path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_learning_review/SKILL.md reference_modules: - id: cold_email_engagement_first_outreach.public_mode_router name: Public Outreach Mode Router summary: Portable public routing guide for choosing the right cold-outreach mode, required inputs, child skill handoffs, and output shape. when_to_load: - When using the portable bundle outside BuildOS. - When the task needs a fast decision between campaign, strategic-account, single-target, investor, recruiting, PR, partnership, or customer-research outreach modes. path: references/public-mode-router.md visibility: public - id: cold_email_engagement_first_outreach.source_map name: Cold Email Source Map summary: Runtime source map for the cleaned source corpus, foundational source layers, tactical gap-fill sources, and local research lineage. when_to_load: - When the agent needs provenance, source conflicts, or links to deeper research documents. path: references/source-map.md visibility: internal - id: cold_email_engagement_first_outreach.internal_os name: Internal Outreach Operating System summary: Internal blueprint for right person -> right moment -> right reason -> right offer -> right ask -> right follow-up. when_to_load: - When designing the broader market-learning and relationship-starting operating system, not just one email. path: references/internal-operating-system.md visibility: internal - id: cold_email_engagement_first_outreach.skill_architecture name: Internal Skill Architecture summary: Internal map for separating root skills, child skills, and tangential relationship-building skills. when_to_load: - When building or reorganizing nested outreach skills. path: references/internal-skill-architecture.md visibility: internal - id: cold_email_engagement_first_outreach.child_skill_source_plan name: Child Skill Source Plan summary: Internal guide for sourcing and deepening the cold email child skills with real source material. when_to_load: - When another agent is assigned to flesh out a cold email child skill with source-backed detail. path: references/child-skill-source-plan.md visibility: internal - id: cold_email_engagement_first_outreach.source_acquisition_queue name: Source Acquisition Queue summary: Internal corpus status and remaining queue for experts, books, sites, videos, and official documents behind the full cold email architecture. when_to_load: - When an agent needs to collect or assign source materials for deepening the cold email architecture. path: references/source-acquisition-queue.md visibility: internal path: apps/web/src/lib/services/agentic-chat/tools/skills/definitions/cold_email_engagement_first_outreach/SKILL.md --- # Cold Email Engagement-First Outreach Root skill for trust-preserving cold outreach. The operating sequence is: ```text right person -> right moment -> right reason -> right offer -> right ask -> right follow-up ``` The north star is qualified conversations started per unit of market trust consumed. ## When to Use - Plan or audit a cold email campaign - Draft one high-value targeted cold email - Choose between high-volume, strategic-account, single-target, reply-revival, investor, recruiting, PR, podcast, partnership, or customer-research modes - Diagnose low opens, low replies, bad-fit replies, poor meeting conversion, objections, or deliverability risk - Design the outreach operating workflow before sending - Decide which child skill should handle a weak input such as ICP, signal, offer, research, taste, deliverability, reply handling, or learning review ## Workflow 1. Choose exactly one mode: high-volume offer test, strategic-account sales, single-target relationship, reply revival, investor fundraising, recruiting, PR/podcast, partnership, or customer research. 2. Check the prime flow in order: right person, right moment, right reason, right offer, right ask, right follow-up. 3. Route to a child skill only if a specific input is missing or weak. Do not load children automatically. 4. Verify sender trust before any scaled sending. If deliverability is unknown, recommend manual or low-volume sending only. 5. Segment volume campaigns to one persona x one narrowing signal. Reject mixed lists. 6. Require a specific research anchor and bridge for strategic, single-target, investor, recruiting, PR, podcast, partnership, and relationship-sensitive outreach. 7. Prefer a useful artifact offer over a meeting ask: note, teardown, benchmark, signal report, sample, snapshot, deck, topic angles, or candidate slate. 8. Draft subject and preview before the body. Reject promotional, vague, money-word, or generic packaging. 9. Draft the body by mode and use one CTA: the smallest useful yes. 10. Add proof only when true, relevant, and approved. 11. Plan cadence by mode and define reply routes before sending. 12. Return a campaign bundle, per-email bundle, audit report, or child-skill routing plan. ## Child Skill Routing - Use `cold_email_icp_signal_design` when the target, persona, timing signal, or disqualifier set is vague. - Use `cold_email_offer_lab` when the offer is missing, meeting-first, too large, or not artifact-shaped. - Use `cold_email_research_anchors` when a high-value email needs a better anchor or bridge. - Use `cold_email_outreach_compiler` when inputs are ready and the user needs the finished bundle. - Use `cold_email_taste_review` when reputation risk, weak specificity, or fake personalization is the main concern. - Use `cold_email_deliverability_readiness` when scaled sending, cold domains, or inbox health matters. - Use `cold_email_reply_os` when replies arrive, silence needs revival, or objections need routing. - Use `cold_email_learning_review` after a test produces metrics, replies, objections, or buyer language. ## Output Contracts - Campaign bundle: mode, segment, sender check, offer, subject, preview, body, proof slot, CTA, cadence, reply routes, tracking targets, and refusal note if needed. - Per-email bundle: mode, anchor, bridge, subject, preview, body, CTA, proof, specificity grade, taste review, and follow-up plan. - Audit report: mode mismatch, sender gap, subject/preview issue, segment/research gap, offer gap, passive language, missing bridge, proof issue, cadence violation, reply-routing gap, and 3 rewrite candidates. - Routing plan: which child skill to load, why, and what artifact it should return to the root. ## Guardrails - Do not send, schedule, or publish outreach without human confirmation. - Do not fabricate research, proof, metrics, mutual contacts, customer names, or buyer-language claims. - Do not run mixed personas in one campaign. - Do not recommend volume sending without verified sender health. - Do not use "worth a chat" as the offer. - Do not use a Loom, deck, or calendar link as the cold value unless the mode specifically supports it. - Do not leave preview text to chance. - Do not use investor meeting-first cold emails. - Do not stack multiple offers or CTAs. - Do not exceed mode cadence limits. - Do not spend market trust casually. ## Examples ### User has a list but no clear segment - Load `cold_email_icp_signal_design`. - Return a segment definition and signal rubric before drafting. ### User has a target and a weak offer - Load `cold_email_offer_lab`. - Return an artifact offer and smallest useful yes before compiling the email. ### User has a drafted email and asks whether it is good - Load `cold_email_taste_review` if the risk is reputation/tone, or `cold_email_outreach_compiler` if the goal is a full rewrite. ## Notes - This is the runtime BuildOS skill. Public blog content may paraphrase it, but should point back to this file as the core skill source. - Deeper research lineage and public source analysis live under `docs/research/youtube-library/skill-drafts/cold-email-engagement-first-outreach/`.