list_agents
The agents your team set up on SocialScout: what each does, when it runs and its last run.
- Returns
- Each agent with its task, schedule, tools, where results go, on or off, and its last run.
- Does
- Reads
- Who
- Paid plan · plans
- Limits
- 1,500 tool calls a day on the paid plan, shared by all tools.
Try it on your account
Your plan does not include this tool. Running it shows the answer your agent would get, which says what it needs.
Give it to your agent
Prompt for
No key in it: your agent reads it from SOCIALSCOUT_KEY or asks you.
Read the prompt
Set up SocialScout for me and use it: Belgian social-media data (viral brand videos, Flemish creators, Meta ads and what they say, the Belgian marketing calendar, my own numbers and my content calendar), as an MCP server.
## 1. Connect
You are Claude Code. Add SocialScout as an MCP server for my user (not in this project's .mcp.json, so the key never lands in git):
claude mcp add --transport http --scope user socialscout https://socialscout.be/api/mcp --header "Authorization: Bearer $SOCIALSCOUT_KEY"
First check that SOCIALSCOUT_KEY is set (echo ${SOCIALSCOUT_KEY:+set}); the shell fills it in when the command runs.
Then tell me to run /mcp (or restart Claude Code) so the socialscout tools load.
My key: use the environment variable SOCIALSCOUT_KEY. If it is not set, ask me for my key (I find it at https://socialscout.be/me/crew, under "SocialScout inside your AI": the part after ?key=). Treat it as a password: never print it back, never put it in a file that goes into git.
What the key may do: it reads everything my plan includes and may make drafts in my calendar (scopes read + drafts). It cannot schedule, publish or touch ads: schedule_post, reschedule_post, unschedule_post and delete_post answer "needs_confirmation" with a review_url. Give me that link; nothing happens until I confirm there.
Check the connection: call corpus_stats (works without a key), then list_accounts (needs the key), and tell me in one line what you see.
## The tool: list_agents
The agents your team set up on SocialScout: what each does, when it runs and its last run.
Returns: Each agent with its task, schedule, tools, where results go, on or off, and its last run.
Does: Reads · Paid plan
Limits: 1,500 tool calls a day on the paid plan, shared by all tools.
Parameters:
No parameters.
## Worked example
"Which of our agents are running, and what did they find last?"
Call list_agents and answer from what it returns, with the links.
## Rules
- Every claim about a video, creator, brand or ad comes from a tool result. Say so when a tool returned nothing.
- Every result carries a link to its page on socialscout.be: put it next to each name.
- "5x" or "viral" means five times the account's own normal views, not a fixed number.
- Times without an offset are Belgian time (Europe/Brussels).
- A tool that answers with a sentence instead of data is telling you what it needs (a plan, a key): pass that on, do not guess.What your agent reads
Paid plan (market research). The team's own agents that SocialScout hosts: what each does, when it runs, its tools, where results go, on or off, and its last run.
More in your numbers
my_watchlistYour private watchlist: each account's last 30 days against the 30 before, and its best videos.watch_accountAdd a TikTok or Instagram account to your private watchlist.unwatch_accountTake an account off your watchlist.my_numbersYour own connected accounts: reach, views, engagement, followers and ad results, against the period before.my_dashboardYour full dashboard history for any dates: medians per platform, a time series, followers, ads per campaign.ask_scoutAsk Scout, the data analyst, anything across the database and your own data, ad reads included.my_brandYour brand profile: company, website, handles, and the brief Scout writes from.my_overviewYour own numbers in short: reach, views, engagement and followers, free on every account.