Find and qualify public buying signals

Lead Signals turns public business activity into a ranked account research list through Treg. A signal is a dated fact that makes a conversation timely, such as a complaint about a relevant problem, hiring for a matching role, funding, a tool change, or engagement with a relevant post.

The workflow separates detection from qualification. Raw signals are filtered for buyer fit before optional contact enrichment, so a large activity list does not automatically become a lead list. Each retained row explains why the account fits and why the timing matters.

Use it for an authorized search for buyers, public market signals, or qualification of supplied company URLs. The default discovery path searches pain-related LinkedIn, X, and Reddit posts, inspects promising comments and reactions, filters for relevant titles, groups by company, and prioritizes fitting accounts.

Business-contact enrichment is optional and restricted to surviving rows. The workflow does not contact people or make hiring decisions automatically. It excludes private personal emails, home addresses, personal phone numbers, and sensitive traits. Installing a catalog does not connect accounts, authorize calls, enable memory, or install Treg.

An approved total budget and per-call caps are required before paid calls. Provider examples and quoted prices reflect a pinned revision; inspect current schemas and prices before use. There are no automatic top-ups or uncapped batches.

Inputs include the user's buyer criteria, requested signal types, permitted enrichment scope, and approved research budget. Search the catalog by the job in plain language rather than a remembered endpoint or vendor:

treg catalog search "<the job in plain words>"

Inspect candidate endpoint details for price, hit rate, and notes. Prefer routed endpoints where available, run one small call, inspect the reported charge and serving provider, then scale within the approved limits. The documented routed email finder is treg.people.email.find, with X-Treg-Route-Max-Cost set to 0.05 to limit routing cost.

Qualification weighs fit first, freshness second, and independent signal strength third. Post research favors recent first-person pain over vendor promotion or generic reactions. Other signals include low-star reviews about the relevant job, recent hiring, fitting job changes, funding stage, technology adoption, buyer-team headcount trends, company news, and active ads.

Filters remove misleading or weak evidence such as evergreen job reposts, staffing firms, intra-group job moves, debt mistaken for growth, acquisition-driven headcount jumps, ubiquitous technology tags, stale reviews, and syndicated news duplicates. Page-change monitoring can combine scraping with a retained hash; custom intent can use topic-specific search.

The result is a ranked list containing person where relevant, company, the signal, a one-line timing reason, source link, and work email only when authorized and found. Report coverage and settled charges. Missing capabilities are reported locally; external catalog requests require explicit authorization.

Recurring monitoring starts with one costed baseline run. A signals.csv file deduplicates later results by person URL, signal, and source URL. Scheduling requires separate recurring-task approval, an authorized token, and a persistent readable and writable location for that file.

The skill works through Treg MCP tools or the treg CLI. Catalog search, catalog detail, and call operations support provider discovery. Existing authorized connections are used, and inputs sent to Treg and its providers are disclosed. Provider availability and live task success remain dependent on the connected service.