Evidence over volume
A shorter list with visible reasoning is more useful than an enormous export with uncertain origin.
Operating principles
Nooks is presented as an agent-native research skill for teams that value traceability, controlled action, and an explicit boundary between a useful inference and a verified fact.
A buyer begins with a falsifiable ICP, not an instruction to “find good leads.” The task records inclusion criteria, exclusions, territory, evidence thresholds, and the intended downstream decision. This makes the research question legible to sales leadership and RevOps before data collection starts.
The agent gathers public company context, role signals, and configured enrichment fields. Sources may disagree or leave gaps. The workflow preserves those gaps instead of silently replacing them with confidence. Provider availability, API rate limits, and data freshness depend on the environment and must be measured locally.
Each candidate should carry a reason for inclusion. The explanation names observed signals, source timing, transformations, and any inference. This evidence trail lets a reviewer challenge the result line by line and prevents a polished interface from becoming a substitute for data provenance.
A human approves targeting, contact choice, and outbound language. Verification does not establish legal permission, guarantee deliverability, or eliminate the need for suppression. Teams remain responsible for CAN-SPAM, CASL, GDPR, CCPA/CPRA, contractual restrictions, and other rules applicable to their operation.
Accepted and rejected results become evaluation evidence. Teams can refine prompts, source priority, and confidence thresholds while monitoring match quality and downstream outcomes. No universal benchmark is asserted: an honest evaluation uses the buyer’s own segments, providers, runtime, and review standard.
A shorter list with visible reasoning is more useful than an enormous export with uncertain origin.
Research, drafting, approval, and sending are distinct states with separate permissions.
When a public source does not confirm a claim, the correct label is unknown—not a guessed answer.
Test whether outbound economics justify another platform or whether an agent-native task fits the existing stack.
Inspect data movement, permission scope, rate limits, export behavior, and reproducibility.
Define review gates that preserve relevance, brand voice, suppression, and sender reputation.
Evaluate package execution, credential storage, provider contracts, retention, and deletion paths.
Disclosure: SOC 2, ISO 27001, GDPR, and related frameworks are evaluation topics, not badges claimed by this site. Confirm any certification or compliance representation using current first-party evidence.
Copy the exact command, run it in a controlled environment, and review each requested permission.
npx -y @okki-global/okki-go-taroball