Repository navigation
A consensus-check tool for Agno Teams (flag when member agents agree for the wrong reasons) #11205
screwtapeai
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
In an Agno Team the leader often combines answers from several members. When members agree, the leader tends to trust it, even if they all leaned on the same weak source. Here's a small custom tool the leader can call before it commits to a combined answer:
Then add it to the team, e.g.
Team(members=[...], tools=[check_member_agreement], instructions=["Before finalizing, call check_member_agreement with each member's answer; if risk is high, ask a member to re-verify against the sources."]).For single-agent drafts,
POST /v1/judgereturnspass,failorretrywith reasons you can feed back to the agent. It's also available as an MCP server if you preferMCPTools.Full runnable example: https://github.com/screwtapeai/screwtape-mcp-clients
Disclosure: I build Screwtape.ai. An independent judge for your agents. Screwtape isn't tied to any agent framework or AI model provider, so its verdicts are never grading their own work. Every verdict comes with a signed receipt you can verify yourself, which gives you a record you can trust when it matters.
judgeandconsensus): https://screwtape.ai/mcpreceiptobject to https://screwtape.ai/v1/receipts/verify.We're looking for a few design partners who run multi-agent teams in production: 1,000 free calls a day for 60 days, plus direct input on the roadmap. Mostly I'd like to hear where this pattern breaks for you. Reply here or email screwtapedev@gmail.com.
All reactions