Sent Agent Skills for SMS, WhatsApp, RCS, and MCP

The official Sent agent plugin combines focused Agent Skills with live Model Context Protocol (MCP) tools. Use it to send and track business messages, manage contacts and templates, query analytics and account readiness, diagnose delivery failures, and guide SMS 10DLC, WhatsApp Business, RCS/RBM, and Sender Profile implementations.

This repository is both:

• a skills collection that can be discovered and installed with the skills CLI; and
• a portable Agent Plugins 1.0.0 package with generated Codex and Claude Code adapters.

Each skill's canonical documentation lives in its SKILL.md. The collection READMEs provide the searchable task index; skill-specific references/, scripts/, and agents/openai.yaml files provide deeper guidance only where needed.

List the available skills without installing them:

npx skills add https://github.com/sentdm/sent-plugin --list

Install the sent dispatcher when you want the agent to route broad Sent requests:

npx skills add https://github.com/sentdm/sent-plugin --skill sent

Install one or more focused skills when the task is already known:

npx skills add https://github.com/sentdm/sent-plugin \
  --skill sent-messaging \
  --skill messaging-performance-analyzer

Install all skills:

npx skills add https://github.com/sentdm/sent-plugin --skill '*'

The Skills CLI installs agent instructions. Live account reads and mutations additionally require the Sent MCP server through a compatible, authorized client.

/plugin marketplace add sentdm/sent-plugin
/plugin install sent@sent

For local plugin development:

claude --plugin-dir ./claude-plugins/sent

The repository root is directly discoverable as an Agent Plugins 1.0.0 package. packages/sent is the canonical portable artifact; plugins/sent and claude-plugins/sent are generated host adapters.

Use sent-analytics for aggregate dashboard totals and trends. Use messaging-performance-analyzer for message-level evidence, funnel drop-off, and root-cause analysis. Use sent-templates for existing records, waba-template-author for WhatsApp content and policy decisions, and template-builder-ui for product UX.

For engineering work, use sent-integration-starter to stand up and harden a new integration, sent-webhook-engineer for receiver construction and delivery debugging, sent-routing-strategist for channel selection and route diagnosis, sent-two-way-messaging for inbound and consent flows, sent-profile-provisioning to execute the profile lifecycle that sender-profile-architect designs, and migrate-to-sent when replacing an incumbent provider.

The plugin declares the Streamable HTTP endpoint https://mcp.sent.dm/mcp and exposes these operations through the authorized connection:

The MCP client performs OAuth 2.1 authorization with PKCE and Dynamic Client Registration. The user selects an organization and Sender Profile during authorization; the client stores the resulting grant. An organization grant can use a schema-supported profileId to act as an owned profile; profile grants cannot. sender_profiles.* uses target id and rejects acting profileId. Reauthorize for another organization or scope outside the grant and revoke access from Sent Dashboard → Settings → MCP Connections. Do not paste API keys or tokens into prompts.

• Preview the exact organization, Sender Profile, target, and payload before every send, contact creation, profile change, feedback report, or deletion.
• Require explicit confirmation immediately before the mutation, including profile changes and feedback reports. Any changed payload or retry needs a fresh preview and confirmation.
• Never retry an ambiguous send automatically; inspect the message and activity history first when possible.
• Treat accepted or queued as processing states, not proof of delivery.
• Mask phone numbers where practical and avoid repeating message bodies, contact data, KYC data, or billing details.
• Treat number lookup as a capability signal, never as evidence of consent.

The canonical source is packages/sent. Root discovery files and host adapters are generated; do not edit generated copies directly.

packages/sent/                # canonical portable package
├── plugin.json               # searchable package metadata
├── mcp.json                  # remote MCP declaration
├── README.md                 # portable-package documentation
└── skills/<name>/
    ├── SKILL.md              # trigger metadata and workflow
    ├── agents/openai.yaml    # optional UI metadata
    ├── references/           # optional deep reference docs
    └── scripts/              # optional deterministic helpers

evals/                        # positive, negative, and overlap routing cases
adapter-sources/              # shared and host-specific adapter sources
plugins/sent/                 # generated Codex adapter
claude-plugins/sent/          # generated Claude Code adapter
scripts/                      # generation, validation, and fixture tests

When adding or changing a skill, update the canonical SKILL.md, keep its frontmatter description rich in user intent and domain terms, add routing evaluations, update the sent dispatcher and README catalogs when scope changes, then regenerate and validate:

python3 -m pip install -r requirements-dev.txt
python3 scripts/generate_adapters.py
python3 scripts/validate.py
python3 scripts/test_validation_gates.py
python3 scripts/test_fixtures.py
claude plugin validate . --strict
claude plugin validate ./claude-plugins/sent --strict

See CONTRIBUTING.md for the public-data policy, docs/PUBLIC_RELEASE.md for the release checklist, and SECURITY.md for security reporting.

• Sent MCP server documentation
• Sent machine-readable documentation index
• Sent API documentation
• Agent Skills specification
• Agent Plugins specification

The public MCP surface includes 27 tools. messages.send uses an existing template with all required parameters; it supports scheduledAt with an explicit timezone offset, from 1 minute to 30 days ahead. Acceptance is not delivery, and quiet hours can defer release.

Feedback goes to Sent staff only with user authorization and a reviewed, sanitized report. It is limited to 2000 characters and 20 calls per authenticated account scope per UTC day, shared across acting profiles. Its note does not guarantee storage, open a support ticket, or promise a response. Number lookup is paid and limited to 1000 calls per authenticated account scope per UTC day.

Profile create and delete require idempotencyKey; interrupted outcomes require reconciliation before another operation. MCP update changes only name, short name, or description. Compliance setup plans create a new profile, support SMS only, and hand required document uploads to the dashboard or REST API.

Dashboard volume counts delivered/read outbound SMS and WhatsApp over supported windows; deliverability is an all-time outbound percentage including SENT, and contacts is a current total. Do not claim unsupported date filters or trends.

Use the connected server's tool schemas for argument names and availability. The public MCP landing page lists the current surface; some documentation pages may describe an earlier catalog.

The bundled migration inventory scanner reads local regular files in the selected repository and makes no network requests. It skips symbolic links and environment files and omits source excerpts from reports. Other bundled utilities validate supplied local payloads or analyze supplied local exports. Remote account reads and authorized writes use the declared Sent MCP endpoint.

The webhook signature utility runs locally with an explicitly supplied signing secret on standard input (--secret-stdin); it makes no network requests and does not read installer credentials or environment variables. Application integration examples require explicitly supplied API keys and signing secrets. Plugin authentication remains client-managed OAuth.