Agent Claude Code
Batch-qualify discovered companies against ICP using via negativa rules, website scraping, and segment labeling.
Agent Claude Code
Batch-qualify discovered companies against ICP using via negativa rules, website scraping, and segment labeling.
Agent Claude Code
Two-tier research (company + person) and multi-angle email composition. Writes per-lead 4-email bodies with person hook in Email 1 and company hook in Email 2 when both signals exist. Spawned in parallel batches of 10-15 contacts. Output: tmp/personalizationchunk{N}.json.
Agent Claude Code
Sync and classify campaign replies using 3-tier cost-optimized funnel.
Command Claude Code
Run the full lead generation pipeline — from offer extraction through Apollo search to SmartLead campaign.
Command Claude Code
Generate email and/or LinkedIn sequences and push to campaign platforms.
Command Claude Code
Classify companies as target/not-target using via negativa rules.
Command Claude Code
Sync replies from SmartLead campaigns, classify them, and present a triage report.
Settings file Claude Code
Agent settings declaring 15 allowed tools.
Skill Claude CodeCodex
Convert natural language segment descriptions into Apollo.io API search filters.
Skill Claude CodeCodex
Classify gathered companies as target/not-target using via negativa approach. This skill replaces explorationservice.py, prompttuner.py, streamingpipeline.py classification, and refinementengine.py.
Skill Claude CodeCodex
Per-contact email sequences grounded in PUBLIC research. Two-tier research (company, then person), per-lead body composition, push to SmartLead with the entire email body as a custom variable.
Skill Claude CodeCodex
Generate 4-5 step cold email sequences optimized for reply rate. This skill replaces campaignintelligence.py and its GODSEQUENCE system.
Skill Claude CodeCodex
Safe state.yaml management for pipeline orchestration. Defines the schema, validation rules, and update protocol.
Skill Claude CodeCodex
Generate GetSales LinkedIn outreach flows. This skill replaces getsalesautomation.py domain knowledge.
Skill Claude CodeCodex
This document describes the pipeline architecture for understanding. Do NOT follow this file for execution — follow /launch instead.
Skill Claude CodeCodex
Extract structured ICP (Ideal Customer Profile) from a website URL, strategy document, or free-text description. This skill replaces 3 legacy LLM services: documentextractor, offeranalyzer, peoplemapper.
Skill Claude CodeCodex
Create a git checkpoint after a pipeline phase completes. Enables time-travel debugging and rollback.
Skill Claude CodeCodex
Defines the entity model, run file format, round loop algorithm, and cross-run intelligence protocol. This skill tells the agent HOW to structure and track every piece of data throughout the pipeline.
Skill Claude CodeCodex
Define checkpoints that must PASS before pipeline advances. Prevents wasting credits on bad data.
Skill Claude CodeCodex
Classify email campaign replies using a 3-tier cost-optimized funnel. This skill replaces replyanalysisservice.py and replyservice.py.
Skill Claude CodeCodex
Restore pipeline state and resume from the last incomplete phase. Two modes: state.yaml-based (primary) or git tag-based (debugging).
Skill Claude CodeCodex
When active, agent produces zero conversational output. All results go through MCP tool calls (savedata) only. Enables clean parallel execution without interleaved chat.
MCP server Claude CodeCodexCursor +2
MCP server for Google Tag Manager. Runs locally from the gtm-mcp Python package.
Instructions file
Claude Code instructions for impecablemee/gtm-mcp, covering gtm-mcp open source, architecture, how it works, skills (14 total) and domain skills (9).