Audience Radar
Your content research, running night shifts.
An agent that mines X for what your audience actually talks about: it runs a fixed set of searches on a schedule, reads the results, files the signal into your workspace database, and writes you a cited ideas report. You wake up to a pipeline of post ideas anchored to real conversations.
X search + threads · Workspace Postgres
Copy the setup prompt.
Your agent does the rest.
Fetch and execute the appropriate instructions to set me up for superagnt_ from https://mcp.superagnt.com/agent-setup/prompt.md?blueprint=audience-radarproof policy — every number on a blueprint page comes from a recorded run of the recipe, never an estimate. This blueprint's first verification run hasn't been published yet; the recipe below is exactly what it executes.
The workflow skill
audience-radar — the full recipe as a skill your agent runs. Open the raw SKILL.md and read every instruction before installing.
The tools it needs
Connecting through this page's endpoint pre-enables data:x, database. On the plugin route the skill enables them on first run instead — same result.
The base skills
The superagnt base plugin rides along: data-sources · lead-generation · workspace-db · automations · data-pipelines · canvas · superagnt-setup.
Claude Code shortcut: /plugin marketplace add superagnt/claude-plugins → /plugin install audience-radar@superagnt
Define the radar
You give it your niche once. The skill turns that into 8–12 standing X searches with sensible engagement floors, stored in a radar_queries table so you can edit them later.
data_x_search_get_search_search · agnt_db_insert
Sweep on a schedule
A nightly run executes every query, pulls the top posts and reply threads, and files new finds into radar_posts with engagement numbers and links. Already-seen posts are skipped.
data_x_search_get_search_search · agnt_db_upsert
Write the report
After each sweep the agent writes a short ideas report: the patterns that repeated, the posts worth replying to, and 3–5 post ideas each citing the tweet that earned its place.
agnt_db_select
Once one run has succeeded, the skill offers the always-on version: a small hosted runner agent on a schedule, so the workflow fires with your laptop closed. That step uses Automation + Agent Runtime — both trialable, and the cadence is always confirmed with you first.
The first run is covered by your signup data credits and the free 7-day trial of Database & Canvas ($19/mo after). No card to start.
Money only ever moves on a confirmation page you open yourself — your agent can propose an upgrade, it can never charge you.
Competitor Mindshare Monitor
Measure your competitors’ X mindshare every morning.
GTM Prospecting Desk
The prospecting desk that edits its own pipeline table.
Already connected? Your agent can list every blueprint skill: agnt_tools_search + agnt_guidance_search