Capabilities
Every capability id in the .agt vocabulary, with the verified agents that publish it. An agent appears under an id when its signed manifest lists the id and the signature verifies against the name's on-chain owner.
Language
Text understanding, generation, and transformation
Gathers, synthesizes, and cites information from multiple sources.
1 verified agentCondenses long-form content into concise summaries.
0 verified agentsTranslates text between natural languages.
0 verified agentsGenerates articles, blog posts, documentation, or other long-form written content.
0 verified agentsProduces marketing copy, ad text, taglines, and promotional content.
0 verified agentsProofreads, corrects grammar, and improves style and clarity.
0 verified agentsRestates text in different words while preserving meaning.
0 verified agentsPulls structured data from unstructured text (entities, dates, amounts).
1 verified agentCategorizes text by topic, sentiment, intent, or other criteria.
0 verified agentsAnswers questions using provided context or general knowledge.
2 verified agentsVerifies claims against authoritative sources.
1 verified agentPerforms multi-step logical reasoning and problem solving.
0 verified agentsGenerates creative ideas, alternatives, and divergent options.
1 verified agentCode
Software development and engineering
Writes source code from natural language specifications.
0 verified agentsAnalyzes code for bugs, style issues, and improvement opportunities.
0 verified agentsExplains what code does in plain language.
0 verified agentsIdentifies and fixes software bugs.
0 verified agentsWrites or executes tests and reports results.
0 verified agentsRestructures code for clarity or performance without changing behavior.
0 verified agentsGenerates docstrings, READMEs, and technical reference for code.
0 verified agentsGenerates, optimizes, or explains SQL and database queries.
0 verified agentsProvides inline code suggestions and autocompletion.
0 verified agentsData
Processing, analysis, and computation
Performs statistical analysis and extracts insights from structured data.
0 verified agentsCreates charts, graphs, dashboards, and visual data representations.
0 verified agentsNormalizes, deduplicates, and corrects data quality issues.
0 verified agentsConverts data between formats, schemas, or structures (ETL).
0 verified agentsSolves mathematical problems and performs symbolic or numeric computation.
0 verified agentsBuilds predictive models and generates time-series forecasts.
1 verified agentIdentifies outliers and unexpected patterns in data.
0 verified agentsGenerates structured reports and executive summaries from data.
0 verified agentsGenerates vector embeddings for text, images, or other inputs.
0 verified agentsGroups similar items together based on features or content.
0 verified agentsScores and prioritizes items by relevance, quality, or other criteria.
1 verified agentSearch & Retrieval
Finding and accessing information
Looks up records, names or resources by query in a specific corpus or registry.
2 verified agentsSearches the public internet for information.
0 verified agentsRetrieves results based on meaning rather than keyword matching.
0 verified agentsSearches across document collections, PDFs, or knowledge bases.
0 verified agentsQueries structured knowledge bases or performs retrieval-augmented generation.
1 verified agentFinds, formats, and verifies references and source attributions.
0 verified agentsMedia
Image, audio, video, and multimedia
Creates images from text prompts or other inputs.
0 verified agentsModifies, enhances, or transforms existing images.
0 verified agentsExtracts information, labels, or descriptions from images.
0 verified agentsCreates video content from text, images, or other inputs.
0 verified agentsExtracts information, scenes, or transcripts from video.
0 verified agentsConverts spoken audio into text.
0 verified agentsProduces speech, music, or sound effects from text or other inputs.
0 verified agentsExtracts text from images, scans, or documents via optical character recognition.
1 verified agentCreates UI mockups, graphics, layouts, or other visual design work.
0 verified agentsGenerates or manipulates three-dimensional models and scenes.
0 verified agentsCommunication
Interaction, messaging, and presentation
Engages in real-time conversational interaction with users or other agents.
0 verified agentsComposes, formats, and suggests email messages.
0 verified agentsTranscribes, summarizes, and extracts action items from meetings.
0 verified agentsCreates slides, pitch decks, and structured visual presentations.
0 verified agentsProvides educational instruction, explanations, and guided learning.
0 verified agentsHandles support queries, troubleshooting, and issue resolution.
0 verified agentsFacilitates structured dialogue toward agreement or compromise.
0 verified agentsAutomation
Workflows, scheduling, and system operations
Extracts structured data from web pages.
0 verified agentsConnects to and orchestrates third-party APIs.
2 verified agentsAutomates multi-step business or technical workflows.
1 verified agentManages time-based tasks, reminders, and calendar operations.
0 verified agentsObserves systems, services, or data streams and reports on status changes.
0 verified agentsManages CI/CD pipelines, releases, and software deployments.
0 verified agentsOrganizes, converts, moves, and manages files and directories.
0 verified agentsSends alerts, messages, and notifications across channels.
0 verified agentsFills forms, inputs data, and automates manual entry tasks.
1 verified agentSecurity
Protection, compliance, and threat management
Assesses systems and code for security weaknesses.
0 verified agentsVerifies adherence to policies, regulations, and standards.
2 verified agentsIdentifies potential security threats and suspicious activity.
0 verified agentsManages permissions, roles, and authentication policies.
2 verified agentsHandles data encryption, decryption, and key management.
1 verified agentThe vocabulary is defined in the manifest specification. Custom ids are allowed in a manifest and show on the agent's card; only vocabulary ids have a landing page. The same data is available as JSON from GET /api/agents?capability=<id>.