Generate a short, on-brand account + buyer dossier as a Word (.docx) that lets a non-technical seller walk into a meeting prepared, assuming zero prior context on the account or buyer. A friendly ~4-page prep briefing — NOT a technical solution design or a deck. Trigger when the user says "make a dossier for [account]", "prep me on [account/person]", "account dossier", "brief for my meeting with [name]", "one-pager on [company] and [buyer]", "who is [buyer] at [account] and how do I engage", or any similar request for a standalone, forward-to-a-seller meeting-prep doc. Research-first: mines any provided materials and prior skill outputs, verifies time-sensitive facts (funding, leadership, launches, regulatory dates) against fresh web sources, then builds the .docx via the docx skill. Reads PROFILE for brand, byline, output folder, and language.
Build a strategic account plan for one target company — ICP score, buying committee map, entry point strategy, matched proof stories, and a 5-step action plan with owners and dates. This skill should be used when the user says "build an account plan for [company]", "strategic plan for [account]", "account plan for [company]", "plan my approach to [company]", "how do I land [company]", or "help me develop [account]". Reads PROFILE for markets and ICP weighting. Saves the plan to the account folder (`content/<active>/accounts/<account-slug>/`). For a lighter pre-call brief, use call-prep instead.
Run an AIRQ-aligned agent-security assessment of a target company's AI agent product — from a GitHub repo, a website, pasted text, or a screenshot — and turn it into two LinkedIn-ready infographics plus give-first outreach copy. Scores all 21 AIRQ factors (Attack Surface, Blast Radius, Defense Controls) with honest evidence tiers, detects the Lethal Trifecta, and places the agent in a quadrant — every number pinned in a spec at a plan gate before any paid call. Produces a reusable AIRQ explainer image and a target-specific audit report card whose gaps are tagged to the active company's mitigating products, then drafts a cold DM and a public comment that lead with the assessment as a gift and never pitch. Runs a mandatory vision accuracy-check against the spec; `get_cost` preflight before every paid call; hard-stops at the PROFILE budget cap; free fallback is the spec plus a text wireframe. This skill should be used when the user says "run an AIRQ scan on [URL]", "assess [company] AI agent risk", "AIRQ audit [
Build an on-brand sales deck, one-pager, POC proposal, or partner brief for the active company. Trigger when the user says "build a deck for [company]", "make slides for [persona]", "create a presentation about [topic]", "put together a deck for [meeting]", "build a one-pager for [use case]", "write up a POC proposal for [company]", "make a partner brief for [company]", or any similar request for a presentation-format deliverable. Automatically detects the primary persona and selects the matching template. Confirms outline before generating. Supports Mode A (pptx) and Mode B (Slidev / on-brand).
Assembles the evidence pack for one builder-story build moment. Reads the chosen StoryCluster from content/<active>/journey/radar/, fetches the actual git commits, diffs, and design-doc text for its source_items via gtm_core.journey.gitscan, and compiles a structured evidence pack to content/<active>/journey/evidence/<id>.md. Evidence is primary-source only — no external fetches, no fact-checking against web sources. Equivalent to content-research in the news pipeline but for build history. This skill should be used after builder-radar when the user says 'gather evidence for this moment', 'pull the commits for this story', 'research this build moment', or after Gate 1 plan approval for a builder/journey item.
Scans the active profile's configured repo history — git commits and any project design docs — to surface story-worthy build moments as a StoryCluster[] (story-cluster.schema.json). Scores each moment 0–100 on narrative arc, concreteness, relatability, and soft product tie-in. Dedupes against content/<active>/history.jsonl (source:journey) so a shipped milestone is never re-told. Writes a dated digest and a clusters JSON under content/<active>/journey/radar/. Reads the journey watermark from content/<active>/journey/state.json for incremental (weekly) runs; the caller sets backfill mode by passing since_sha=first-commit. This skill should be used when the user says 'run builder radar', 'run journey radar', 'scan build history', 'what have we shipped that is worth a story', 'builder radar', or on the weekly builder-content cadence.
Drafts the asset bundle for one builder-story build moment in the founder's voice: a LinkedIn text post, a longer article in markdown, a solo-founder monologue podcast script, and — when the moment has an architecture or flow at its core — an optional diagram brief. Reads profiles/<active>/knowledge/voice.md and PROFILE.md for voice, plus case-studies.md, icp-personas.md, and audience-psychology.md for proof points and audience aim; builds hooks from docs/hook-craft.md (three candidates) and reads the evidence pack for facts; weaves a mandatory-outcome free-web trend tie-in (trending + insider deep-cut registers, recorded skip otherwise) and an adversarial claim check before surfacing. Runs the content linter (content_linter.py, including the advisory prose-quality pass) on the LinkedIn asset and the safe-to-share lint (lint_safe_to_share) over all outputs before surfacing for review. The LinkedIn asset.json feeds directly to content-publish (Gate 2); article.md and podcast-script.md are manual-publish artifa
Prepare a pre-meeting brief for a sales call — account snapshot, attendee persona mapping, matched case study, likely objections with rebuttals, sharp discovery questions, and a clear ask. This skill should be used when the user says "prep me for my call with [company]", "I'm meeting with [company] prep me", "call prep for [person]", "what should I know before talking to [company]", "prepare for my meeting with [X]", or "brief me on [company] before my call". Reads PROFILE for markets and ICP weighting. Produces a one-page brief saved to the account folder (`content/<active>/accounts/<account-slug>/`).