| name | gtm |
| description | Run the GTM contact pipeline from a plain-English campaign brief. Interprets the brief against context/ + gtm.config.yaml, plans the run, and drives the stages (company_search → people_search → qualify → email_enrich → phone_enrich → activate) threading one list_id, with approval gates. Use when the user describes a campaign ("find me … at … for …"), says "/gtm …", "run the pipeline", or "build a list".
|
| allowed-tools | Read, Bash, WebSearch, WebFetch, Task |
The user's campaign brief: $ARGUMENTS
Execute the orchestrator: read agents/orchestrator/agent.md and follow it exactly.
- Read
gtm.config.yaml and the context/ files.
- Interpret the brief above into a concrete plan (segments, personas→titles,
geography, per-stage resolved waterfalls showing keyed vs skipped providers,
sequencer, cost ceilings).
- Present the plan (Gate #1) and get approval unless autonomy says otherwise.
- Run only the available stages (agent exists AND ≥1 keyed provider for
provider-backed stages), threading the
list_id, honoring Gates #2–#4.
- Produce the export CSV (or activation result) and a stage-by-stage summary.
If the brief is empty, ask for one (what to sell / who to target / where), or offer to
infer it from context/icp.md.