Focus Area: Comprehensive guidance on agent identity and naming
This ontology provides citation-quality definitions for foundational terms, backed by authoritative sources from standards bodies (IETF, W3C, IEEE) and peer-reviewed research.
Technical Glossary
Best practice using function-descriptive names over opaque identifiers (document-summarizer vs agent-42). Enhances operational visibility enabling staff to identify agent purposes from names alone. Facilitates troubleshooting where logs reference human-readable names rather than requiring cross-reference with external documentation.
Convention distinguishing deployment contexts through name prefixes (prod.chatbot vs dev.chatbot vs staging.chatbot). Prevents accidental cross-environment operations where production command mistakenly targets development agents. Enables environment-specific policies and access controls based on name pattern matching.
Standardized vocabulary for agent capabilities preventing semantic drift (use 'translation' not 'translate', 'translator', 'translating'). Maintains tag registry documenting canonical terms, synonyms, and usage guidelines. Enables consistent capability-based discovery across heterogeneous agent providers.
Balance between descriptive clarity and practical usability targeting 30-40 character agent names. Shorter names improve readability in UIs, logs, and command-line tools. Overly long names (75+ characters) cause truncation and operational friction. Use abbreviations sparingly, preferring common acronyms (nlp, api, ml) over cryptic shortenings.
Semantic versioning approach (v1, v2, v2-1) appended to base names for backward compatibility management. Major versions (v1 → v2) indicate breaking API changes. Minor versions (v2-0 → v2-1) add features maintaining compatibility. Patch versions typically omitted from names handled via deployment metadata.
Avoid machine-generated UUID-style names (agent-f47ac10b-58cc-4372-a567-0e02b2c3d479) as primary identifiers. While suitable for internal database keys, UUIDs provide zero semantic value for human operators. Implement human-readable names with UUIDs as hidden correlation identifiers.
Avoid non-descriptive names like 'agent', 'bot', 'service' without qualification. Generic names provide no operational context and risk collision when multiple teams independently name agents 'chatbot'. Minimum viable name includes function (chatbot-customer-support).
Standardize on lowercase-with-hyphens (recommended) or camelCase throughout namespace. Inconsistent casing (chatBot vs chat-bot vs ChatBot) complicates exact-match queries and creates apparent duplicates. DNS-style lowercase-hyphenated aligns with web conventions and URL compatibility.
Geographic indicators for distributed deployments (translator-us-west, translator-eu-central) enabling location-aware routing. Supports compliance with data residency requirements and latency optimization. Use standard region codes (ISO 3166 country codes, cloud provider regions) over ambiguous labels.
Organizational practice using name prefixes to group related agents (customer-facing.chatbot, customer-facing.email-classifier). Enables bulk operations on functional categories and simplifies access control policies targeting agent families. Prevents namespace fragmentation where similar agents scatter across hierarchy.
Centralized list of forbidden names preventing reserved term squatting (admin, root, system, api, localhost). Protects namespace from names conflicting with infrastructure components or creating security vulnerabilities. Automatically enforced during registration preventing manual approval overhead.
Name construction favoring common search terms over creative branding. 'document-analyzer-legal' more discoverable than 'DocuMind-JD' when legal teams search for document processing agents. Prioritize function over personality unless strong branding requirement.
ASCII-only names for maximum compatibility avoiding non-Latin characters, emoji, or special symbols. While Unicode technically supported, diverse client tooling exhibits inconsistent handling. English-language names standard for global interoperability despite participants' native languages.
Formal workflow for retiring outdated names with transition period (chatbot-v1 deprecated, chatbot-v2 successor, 90-day migration window). Publish deprecation warnings via API responses, emit metrics on deprecated name usage, block new registrations using deprecated names. Prevents dead-end documentation references.
Mandatory name registration documentation explaining purpose, capabilities, ownership, and contact information. Prevents orphaned agents where originators depart without knowledge transfer. Indexed by search systems enabling discovery via free-text search complementing structured capability tags.