Install the Skill
Add the SocQ Skill from its public repository to a compatible coding agent.
npx skills add SocQAPI/socq-devtoolsA public social data research Skill built for AI agents.
Keep your existing MCP setup. Add one Skill to help your agent choose endpoints more accurately, control cost, and reliably complete asynchronous, paginated, and cross-platform research.
One Skill for the complete SocQ API catalog12 supported platforms92 API capabilities
Quick start
Three steps to give your agent a complete research workflow
Install the SocQ Skill, configure your API key securely, and describe the research goal in natural language. Your agent follows the Skill to select endpoints, execute tasks, paginate results, and organize the final analysis.
Configure your API key
Store the key in a local environment variable so live SocQ tasks can authenticate safely.
SOCQ_API_KEYDescribe the research task
Ask in natural language. The Skill guides endpoint choice, execution, polling, pagination, and result organization.
Compare recent posts from these three X profiles.Never paste an API key into a public chat, repository, or shared document. Follow the client-specific guide for secure local configuration.
Read the complete install guideMCP And MCP + Skill
MCP vs. MCP + Skill
MCP gives agents direct access to SocQ’s public social data tools. Add the Skill to improve endpoint selection, estimate cost, complete asynchronous and paginated tasks, and run cross-platform research with a consistent workflow.
MCP only
Direct access to live SocQ tools
MCP already lets an agent discover endpoints, submit collection tasks, inspect task status, and read results.
- Call a known API with a clear input
- Submit and read a simple collection
- Let the model decide the surrounding workflow
Best for known endpoints and straightforward one-off collections.
MCP + Agent Skill
The same tools with a reliable research playbook
The Skill teaches the agent how to choose, sequence, and recover SocQ operations without adding a separate data service.
- Map ambiguous intent to the right endpoint
- Check schemas and cost before large jobs
- Poll, paginate, retry safely, and normalize research output
Best for multi-step, costly, asynchronous, repeated, or cross-platform research.
| Task | MCP only | MCP + Skill |
|---|---|---|
| Call a known endpoint | Call the tool directly and read the results | Validate inputs first, then execute and organize results with a consistent workflow |
| Turn a research goal into API steps | The agent chooses endpoints and execution order case by case | Apply a repeatable method to break down the goal, select endpoints, and plan each step |
| Run a large asynchronous collection | The agent handles cost, task status, and pagination case by case | Estimate cost before execution, continue the original task, and read every requested page |
| Run a cross-platform comparison | The agent defines the collection and comparison approach each time | Standardize limits, collection steps, and output structure so results are directly comparable |
If you only call a few clearly defined SocQ APIs, MCP is enough. Add the Skill when you want the agent to independently complete multi-step, asynchronous, or cross-platform research.
Explore SocQ MCPWhat the Skill teaches
Guide agents from a research goal to complete results
The SocQ Skill provides a reliable method for endpoint selection, input validation, cost estimation, asynchronous tasks, pagination, and error recovery—making public social data research more accurate, efficient, and reusable.
Choose the right capability
Map natural-language intent to the relevant endpoint in the SocQ Capability Registry.
Shape valid inputs
Determine whether a task needs URLs, usernames, hashtags, queries, or another documented input.
Control task cost
Inspect the capability and estimate credits before submitting a large collection.
Complete asynchronous work
Continue polling the original task instead of resubmitting and potentially paying twice.
Follow pagination and files
Read cursor-paginated results and retrieve raw output files when the workflow needs them.
Recover from failures
Respond to authentication, validation, rate, credit, and provider errors with the appropriate next step.
Research prompts
Describe the research goal. Let the Skill plan the workflow.
Each example starts with a business outcome. The Skill selects endpoints, validates inputs, controls cost, completes asynchronous tasks and pagination, and organizes the results into analysis ready to use.
Instagram Posts API
Uses: instagram/postsTikTok Search API
Uses: tiktok/searchTwitter Search API
Uses: x/searchYouTube Search API
Uses: youtube/searchLinkedIn Profiles API
Uses: linkedin/profilesReddit Posts API
Uses: reddit/postsHow it works
From a research question to structured results
The Skill adds a decision layer around the same SocQ capabilities available through MCP, CLI, and REST.
Read the complete Skill workflowUnderstand the goal
Identify the entities, platforms, time range, result volume, and desired output.
Select endpoints
Match the task to current SocQ Capability Registry entries.
Inspect inputs and cost
Check schemas, constraints, and estimated credits before a large job.
Execute the collection
Prefer MCP and pass the Skill attribution marker when submitting.
Continue the task
Poll the returned task ID instead of creating duplicate collections.
Read every requested result
Follow cursor pagination and retrieve files when needed.
Organize the answer
Return comparable tables, summaries, evidence, or structured output instead of raw endpoint jargon.
Execution preference
Hosted MCP → SocQ stdio bridge → CLI or REST fallback when MCP is unavailable.
Why SocQ Agent Skill
Give agents a reliable method for public social data research
The SocQ Agent Skill stays aligned with the live Capability Registry, helping agents choose endpoints accurately, complete asynchronous tasks, plan cost, and compare platforms with a consistent methodology.
01
Aligned with the Capability Registry
Endpoint selection and input guidance reference the same capability source used by MCP and the API catalog.
Skill ↔ Capability Registry ↔ MCP02
Native asynchronous workflow
The agent knows when to submit, when to wait, and how to continue without duplicating work.
submit → task_id → poll → results03
Cost-aware execution
Large tasks can be checked and planned before they consume credits, with idempotent retries when appropriate.
schema → estimate → execute04
Cross-platform research discipline
Use consistent collection limits and output structures when evidence comes from different platforms.
one goal → multiple platforms → comparable outputChoose an integration
Where should your workflow connect to SocQ?
MCP, CLI, and REST are execution paths. Agent Skill is a guidance layer. They share one API catalog and asynchronous task model, so you can choose the interface that fits your workflow.
Current page · Agent Skill
Give agents a reliable method for data research
Add endpoint selection, cost control, pagination, and task recovery guidance. Pair it with MCP first, or execute through CLI or REST.
- Best for
- Multi-step, asynchronous, and cross-platform research
- How it works
- Agent Skill + MCP (preferred)
- Next step
- Install SocQ Agent Skill
Preferred execution path
MCP
Connect once, then let agents discover tools in natural language, submit collections, and continue through asynchronous task results.
Agent Skill FAQ
Frequently asked questions
What is the SocQ Agent Skill?
It is a reusable set of instructions that teaches an AI agent how to select SocQ endpoints, shape inputs, control cost, complete asynchronous collections, and organize research results.
MCP already works. Why install the Skill?
MCP gives agents access to live SocQ tools; the Skill gives them a reliable method for completing end-to-end research. With the Skill, agents translate goals into steps, select endpoints, validate inputs, estimate cost, continue asynchronous tasks and pagination, and organize cross-platform results with consistent criteria. For multi-step, long-running, or repeatable research, this reduces ad hoc decisions and duplicate execution while producing more reliable, complete, and comparable results.
Does the Skill unlock more APIs than MCP?
No. The Skill does not add a separate data service or unlock private capabilities. It teaches the agent how to use the same SocQ API catalog more reliably.
Do I have to install MCP to use the Skill?
No. The Skill prefers hosted or stdio MCP, but it can guide the agent to use the SocQ CLI or REST API when MCP is unavailable.
Is SocQ one Skill or a separate Skill for every platform?
SocQ uses one cross-platform Skill. The agent selects the appropriate capability from the complete SocQ catalog according to the task.
Does the Skill require an API key?
Live data collections require a valid SocQ API key. The Skill itself should never contain or store credentials; configure the key securely in the local client environment.
Can the Skill handle asynchronous tasks and pagination?
Yes. It guides the agent to continue polling the original task, follow result cursors, and retrieve raw files when the requested workflow requires them.
What data can the Skill access?
It can use the public social data capabilities currently available in the SocQ API catalog. Each API page documents supported inputs, public data scope, and normalized result fields.