LeadGrowGTM

Full Attio CRM CLI for AI agents. 20+ command groups covering records, notes, tasks, comments, threads, lists, list entries, objects, attributes, files, meetings, webhooks, workspace, and interactive REPL. Use when you need to read or write Attio CRM data from an agent or terminal. Supports --json on all commands for machine-readable output. Supports --yes on all delete commands for non-interactive use.

Pulls a contact's Attio record + most recent call note via attio-cli, then generates a personalized follow-up email in LeadGrow voice grounded in what was actually discussed. Interactive review loop — show draft, wait for approval or edits, then create an Attio task ("send follow-up to [name]") with a 24h deadline and optionally push the draft to EmailBison. Supports single contact mode and --bulk mode to batch all call notes from the last N hours into a review loop. Trigger when Mitch says "write a follow-up for [name]", "post-call email for [contact]", "follow up on my call with [name]", "write follow-ups for all my calls today", or similar.

One-command pipeline health check. Pulls all active deals from Attio, calculates days inactive from last note, groups by stage, flags stale/cold/dead deals, and renders a ranked summary table with total pipeline value and stage breakdown. Use for weekly reviews, Monday morning check-ins, or anytime you need the full picture without drilling into individual deals. Supports --stage to filter by pipeline stage, --stale-days to customize the inactivity threshold, and --export to save the snapshot to a file. Triggers: show me the pipeline, pipeline snapshot, pipeline health, pipeline overview, how's the pipeline, what's in the pipeline, deal overview, weekly pipeline review, Monday morning pipeline, pipeline status, pipeline report.

Pull the full context on any Attio deal — stage, linked contacts, notes, open tasks — then diagnose where it's stuck and recommend the exact next action with a drafted message or talking points. Interactive review loop: Mitch approves, edits, or redirects before anything writes back to Attio. Supports --all-active for full pipeline review with a triage table first, --bulk-approve to fast-path obvious actions (overdue tasks, proposal follow-ups) without reviewing each deal individually, and --export to generate all deal briefs to a file. Use during weekly pipeline review, before calls, or when a deal has gone quiet. Triggers: deal review, pipeline review, next step, deal stuck, what should I do with this deal, follow up on deal, pre-call prep.

Interactive nurture workflow for stale pipeline contacts. Pulls contacts from an Attio list or stage (e.g. "Nurture", "Interested - No Response"), filters to those with no activity in the last N days, loads LeadGrow voice + ICP context, and generates a tailored follow-up message per contact based on their history, role, company, and last touchpoint. Presents contacts one at a time for Mitch to approve, edit, or skip. On approve: queues message to EmailBison and creates a follow-up task in Attio. Supports --bulk-approve (approve all without per-contact review), --export (generate all drafts to file, nothing sends), and --multi-list (run across multiple lists in one session). Use this when pipeline is going cold, before weekly reviews, or when a list has 5+ contacts with no recent touch.

Pull all overdue Attio tasks, fetch each contact's full note history, generate a personalized follow-up message in LeadGrow voice, and present one at a time for Mitch to approve, edit, skip, or mark complete. On approval, updates the task deadline and optionally queues the message to EmailBison. Supports --bulk-approve to approve all drafts without per-contact review, and --export to generate all drafts to a file without sending anything. Use this during morning pipeline review, weekly pipeline cleanup, or any time you hear "who have I been ghosting" or "what follow-ups are overdue."

Generate Step 1 + Step 2 cold email sequence from ICP research output. Use when user asks to write a cold email, sequence, or outreach copy for a company. Requires ICP research doc or company context.

Research a company and score ICP fit. Use when user asks to research a company, qualify a prospect, or assess ICP match before writing a sequence. Requires domain or company name.

Write cold outbound emails for a specific ICP. Accepts company context, pain points, and sequence position. Returns subject line + body ready for Bison upload.

Research a company and score ICP fit. Takes company name/URL, returns structured fit report with trigger signals, pain hypothesis, and recommended sequence angle.

Encode sales training into a pre-call brief. Takes prospect context + your sales framework (transcript, doc, or notes). Returns a structured discovery call prep sheet with angles, objections, and opening questions.

Build validated web research processes through self-annealing loops. Takes a research goal, generates search steps, tests against sample companies, scores accuracy, and iterates until 90%+. Use when creating new research workflows, building claygent/agent prompts, or systematizing any web research task.

Build validated web research processes through self-annealing loops. Takes a research goal, generates search steps, tests against sample companies, scores accuracy, and iterates until 90%+. Use when creating new research workflows, building claygent/agent prompts, or systematizing any web research task.

Standardized web research agent using FireCrawl (deep scrape) + SerperDev (search) + gpt-4o-mini (extraction). Validated against ground truth with annealable scoring loop. Trigger: any research task that needs web-sourced structured data extraction.

Graduate a research process (processes/find-*.md) into a production trigger.dev scheduled task. Use when user decides a monitoring process is ready for automated daily/weekly execution. Handles config generation, task scaffolding, deploy, and validation.

Run the prompt optimization anneal loop for any scenario. Analyzes failures, generates mutations subject to diversity constraints, executes prompts, judges outputs, and tracks state. Enforces METHODOLOGY.md rules to prevent example-overfitting. Use for any scenario under auto-prompt-creator/scenarios/.

Optimize an underperforming prompt using the anneal loop. Takes a prompt that produces inconsistent or low-quality outputs on a cheap model (Haiku), runs iterative optimization with mutation diversity constraints, and graduates a prompt hitting 92%+ accuracy. Works as guided flow (human-in-the-loop) or autonomous (subagent). Invoke with /anneal-prompt or say "anneal this prompt."

End-to-end account list building using DiscoLike CLI. From seed domain or ICP description through validated, normalized prospect lists ready for Clay enrichment. Includes validation loop, market mapping, and campaign brief integration. Use when building prospect lists, finding lookalike companies, sizing a market, or sourcing target accounts for outbound campaigns.

Search and enrich contact and company data via AI Ark CLI. Interactive intake flow, three-tier title-based contact search (no seniority filter), dry-run URL verification, and domain paste file generation for batch searches. Use when building prospect lists, enriching domain lists with contacts, finding fresh leadership hires, or running targeted people searches.

Manage Trigger.dev background jobs, runs, schedules, and queues from the terminal