Generate outbound email drafts at scale
outbound-pipelineskillsetup L3★1
SRKrukowski/outbound-pipeline-playbook ↗What it does
Generate outbound email drafts from ICP specification end-to-end
Best for
Sales and founder-led sourcing when you have a clear ICP and want to batch-generate cold email drafts without manual research per prospect.
Inputs
- · icp-spec.md file with ICP definition
- · Limit parameter (how many prospects to process, default 10, max 50)
- · Optional dry-run flag
Outputs
- · Count of candidates surfaced by Exa
- · Count of enriched candidates (verified email + title via Apollo)
- · Count of drafts generated by Claude
- · Count of staged Gmail drafts (or JSONL path in dry-run)
- · Failure messages with specific stage and error
Requires
- · Exa (search/discovery)
- · Apollo (email verification + title enrichment)
- · Claude (draft generation)
- · Gmail API or Composio (draft staging)
- · Python orchestrator script
Preconditions
- · icp-spec.md exists in working directory or specified path
- · Exa API credentials configured
- · Apollo API credentials configured (can be skipped with DRY_RUN=true)
- · Gmail credentials configured (can be skipped with DRY_RUN=true)
- · Python environment with dependencies installed
Failure modes
- · Missing icp-spec.md with no user guidance
- · Apollo enrichment fails silently if credits exhausted
- · ICP query too broad → Exa returns hundreds of irrelevant results
- · Draft quality degrades if ICP lacks clear value prop definition
- · Gmail drafts stage fails if permissions incorrect
- · Dry-run mode skips enrichment and staging (output is local JSONL only)
Trust signals
- · Explicit hard rule: never auto-send emails (always drafts for review)
- · Never writes to Sent folder (only Drafts)
- · Asks user to run playbook walkthrough if ICP file missing
- · Reports specific missing env vars and stops (no fallbacks)
- · Common adjustments documented (shorter drafts, different audience, run more prospects, stop using Apollo)