Cold Outreach Command Center — Treg + Apify + IONOS / Cal.com

PIPELINE STATUS

6-Step SDR Pipeline — Active (Permission-Only Cost APIs)

NOT GATED ANYMORE. Cost-bearing APIs (treg `/call/` categories: lead_enrichment, image_generation, video_generation, ai_generation; IONOS SMTP email; Apify budget >0) are ACTIVE but will ONLY run with explicit user request / permission — agent asks before any spend call. Cal.com scheduling and non-metered services remain open. Default budget display shows 0 for visibility, but agent must confirm with user before executing any cost call.

Apify — ACTIVE (limit:5; ask before spend) treg — ACTIVE (ask before any cost call) IONOS SMTP — ACTIVE (max 10/day; ask before send) Cal.com Scheduling — ACTIVE (non-metered) Direct Enrichment — BLOCKED

BUDGET CONTROL

TREG + APIFY Budget — Active With Permission Gate

ACTIVE — COST APIS REQUIRE USER REQUEST / PERMISSION: `TREG_BUDGET_LIMIT` and all `treg.json` categories (`lead_enrichment`, `image_generation`, `video_generation`, `ai_generation`) are active but agent MUST ask user and get confirmation before any `/call/`. `IONOS` email (`OUTREACH_MAX_PER_DAY`) defaults 10/day defensive but requires user approval before sending. Only `Cal.com` scheduling runs without permission (non-metered). No cost call executes silently.

Env vars (inline — never literal embedded):
TREG_TOKEN=${TREG_TOKEN}
TREG_BUDGET_LIMIT=${TREG_BUDGET_LIMIT:-0}
IONOS_SMTP_HOST=smtp.ionos.com
IONOS_SMTP_USER=stephen@citylineexpress360.com (configurable via OUTREACH_FROM_EMAIL)
IONOS_SMTP_PASS=${IONOS_SMTP_PASS} (masked reference — `.secret/ionos_smtp` or env; NOT embedded in SKILL.md/.mcp.json per taste preference)
OUTREACH_MAX_PER_DAY=10

PIPELINE FLOW

Cold Outreach — 6 Steps (Gated)

The pipeline runs from `treg` SKILL.md + `apify-engagers` SKILL.md. Step 3 (`Apify` — `POST_REACTIONS` actor) extracts engagers. Step 4 (`treg` enrichment proxy) enriches via audited `/call/<catalog-id>` endpoints (`apollo/people/match`, `hunterio/combined/find`, `pdl/v5/person/enrich`). Step 5 (`LLM` personalization — custom AI emails using enriched data + ICP + client's `pages.dev` site research + Hormozi-style $100M Offer template from user's 10-lead demo messages). Step 6 (`Cal.com` — scheduling / appointment lock after positive reply). Scheduling is NOT cost-metered.

Pipeline: Apify(3, gated) → treg(4, gated) → LLM(5, separate cost) → Cal.com(6, scheduling/non-metered). Budget: TREG_BUDGET_LIMIT=0 (block); APIFY_ENGAGERS_BUDGET_LIMIT=5; IONOS email max 10/day. Only Apify + treg + IONOS active per user.

GATED CATEGORIES

Cost-Bearing Categories — All OFF by Default

Per `treg.json` (`line 90-96`) and `apify-engagers` budget disclosure (`line 170-176`):

lead_enrichment (`enabled: false`, `limit: 0`, max $0.0015/call — `apollo`/`hunterio`/`pdl` via treg `/call/`)
image_generation (`enabled: false`, `limit: 0`, max $0.015/call — `meigen` / `higgsfield` via treg)
video_generation (`enabled: false`, `limit: 0`, max $0.56/call — `higgsfield/video` via treg)
ai_generation (`enabled: false`, `limit: 0`, max $0.10/call — LLM reasoning/personalization)
email_outreach_ionos (`enabled: false`, `limit: 0`, `max_emails_per_day: 10`, `IONOS_SMTP` — defensive cap vs spam/block)
scheduling (`enabled: true`, `metered_by_treg: false` — `Cal.com`; NOT metered)

Agent must ask: "Treg budget for [category] is 0 (default OFF). Set TREG_BUDGET_LIMIT (or sub-limit) > 0 to approve spend. Proceed?" / "IONOS email gate at 0/day (max 10/day defensive to avoid spam/block for citylineexpress360.com). Budget approved? Proceed?" (`line 127` ask message; `line 175-176` IONOS gate; `line 145-147` scheduling separate).

INTEGRATION

Connected Skills + Deploy Site

Skills Stack

`treg` (agent-tool proxy, budget-gated)
`apify-engagers` (engager extraction — kept separate per user instruction)
`monid` (data discovery — treg proxy alt)
`ai-slop-defeat` (treg layer note added)
`tastemaker` (`treg_proxy:` option in `mcp:` block)
`higgsfield` (budget-gated imagery/video; treg can proxy)

Live Site

Deployed: sdr-neon-guide.pages.dev (`#big` anchor)

`sdr-final.html` pipeline (6 steps) preserved. `Cal.com` scheduling (`Step 6`) — appointment lock after positive reply; scheduling cost tracked separately (NOT through `TREG_BUDGET_LIMIT`).

EVIDENCE / NO HIDDEN SPEND

Verification Gate — Evidence Before Deploy

Evidence (file:line citations — verified this session with `read_file`):

`treg/SKILL.md:2` — name: treg; `line 4`: version 1.0.0; `line 14`: budget-metered tag; `line 16`: `mcp_url`
`treg/SKILL.md:107` — `GATED CATEGORIES` paragraph; `line 109-127`: per-category gates (`enabled: false`, `limit: 0`) + agent ask message; `line 85-105`: `treg.json` skeleton (`default_credits: 0`); `line 95`: `email_outreach_ionos` (`enabled: false`, `max_emails_per_day: 10`, `IONOS_SMTP_HOST`, configurable `OUTREACH_FROM_EMAIL`); `line 96`: `scheduling` (`enabled: true`, `metered_by_treg: false`)
`apify-engagers/SKILL.md:170-176` — budget disclosure (`Apify` ONLY enabled; `treg` proxy; `LLM` separate; direct enrichment GATED OFF; `IONOS` SMTP gate — max 10/day defensive; scheduling separate/non-metered; session-spend confirmation — "no hidden/unmentioned spend")
`.mcp.json` (`python -c` JSON parse confirmed this session) — `treg` entry (`streamableHttp`, URL `https://treg.to/mcp/`) + `higgsfield` (`:-0`) + `meigen`/`monid`/`retell`/`firecrawl` present; `.herenow.env` masked (`# TREG_TOKEN` — no literal); `.secret/` empty (user provides live token); memory (`treg-skill.md`) saved.
`PSTACK AGENT` (`pstack-rigor` SKILL.md — 387 lines, feature/debug/autonomous/interrogate/playbooks; `subagent_type: "pstack-how-critic"` used; `subagent_type: "pstack-agent"` unavailable locally per upstream docs `line 361`; manual verification completed — no hidden spend confirmed).

GATE STATUS Default: ACTIVE — permission gate active. Cost APIs only run when user explicitly requests / approves. No live executions performed without user say-so. User activates `.secret/ionos_smtp` + confirms budget approval when ready. `.backup-cold-outreach-cc-...` preserved.