sales-bizdev

Router skill: resolve ONE topic, Read EXACTLY ONE playbook from routes:, execute its spine. [DOC]

When to use

Any sales/bizdev deliverable: cold/warm outreach, account dossiers, prospect/lead lists, executive pitches, proposals, or sales collateral. Not for delivery, billing, or post-sale account management. [INFERENCIA]

Routing

1. Map the request to one enum topic (above). Pick by deliverable, not channel. [INFERENCIA]
2. Tie-break overlaps: list of accounts → client-prospecting; scored/qualified pipeline → lead-generation; one-account deep brief → client-dossier; per-contact message → b2b-outreach. [SUPUESTO]
3. If two topics fit equally or none does, ask one disambiguating question — never load two playbooks. [DOC]
4. depth=deep → apply the playbook exhaustively, verifying at each step; quick → essentials only. [DOC]

Inputs / Outputs

• In: topic, depth, plus the request's account/ICP/context. Missing target account or ICP for prospecting/dossier work is [SUPUESTO] — fill or ask, don't fabricate. [INFERENCIA]
• Out: the deliverable defined by the chosen playbook, evidence-tagged. Estimates follow Constitution P8 (computed from decomposition + sources, never guessed). Under a commercial profile that forbids client-facing prices (e.g. MetodologIA), express effort in units (FTE-months) + disclaimer, never an unauthorized quote; a profile that permits pricing may state figures. [DOC]

Guardrails (Alfa core tags only — kit-facing)

• Tag non-obvious claims with [CÓDIGO]/[CONFIG]/[DOC]/[INFERENCIA]/[SUPUESTO]; one tag, one spelling, one family per output. See references/verification-tags.md. [DOC]
• Script-first: prefer a repo script over ad-hoc steps where one exists. [DOC]
• Single-brand: identify the brand first; never mix brands in one output. [DOC]
• Quality gates: constitution v6.0.0 (enforcement) + evidence tags. [CONFIG]
• Routing, scoring, and gate aids live in assets/ (routing-matrix.json, quality-rubric.json, preflight-checklist.md). [CONFIG]

Acceptance criteria

• Exactly one playbook loaded; topic ∈ enum; cluster never bulk-loaded. [DOC]
• Output matches the playbook's spine (Discover → Analyze → Execute → Validate) and depth. [DOC]
• Every non-obvious claim carries one Alfa-core tag; no invented prices; single brand. [DOC]
• Each [SUPUESTO] pairs with a concrete next step that would verify it. [DOC]

Anti-patterns

• Loading multiple playbooks "to be safe", or answering inline without reading the route. [INFERENCIA]
• Guessing topic on a genuine tie instead of asking. [INFERENCIA]
• Emitting a price, mixing brands, or using Jarvis-set {…} tags in this kit output. [DOC]

Self-correction

Two routes feel right → re-read tie-break rules; still tied → ask. About to state a number → tag and check it's not a price. Mid-task you find the wrong topic was chosen → stop, re-route, restart the spine. [INFERENCIA]