Skip to content
NEWMarketplace and the trig8n Agent: sell what you build, and get plans and changes that are tested before you see them. See what is new

Library / AI Agents & Assistants

AI Agents & Assistants

Inspect Agentel MCP Skill evidence and metering with the MCP Client

When you run it, it runs an AI agent.

What it does

# Inspect Skill evidence before adoption ### How it works Run this manual workflow to inspect one public Agentel Skill record before deciding whether to adopt it. It checks the Reader identity, current prices and allowance, attempts one detail read only when its local checks allow it, and compares the before/after allowance snapshots. A refused preflight uses the Read not started branch. Reading does not install or execute a Skill, grant permissions, apply to a Mission or award Credit. ### Setup Obtain a restricted Agentel INFO_READ Reader, never an owner or full runtime key. Store it in a private n8n Bearer Auth credential and bind the same credential to all three MCP Client nodes. Edit only skill_id in Choose Skill. After import, open More actions → Settings: set all four execution/progress save options to Do not save, enable a 90-second timeout, save and reopen to verify. Import may discard these settings. Check any automatically selected credential before running. At validation, the detail cost 1 INFO of plan allowance, not an automatic card charge. Recheck current prices; preflight is not a reservation. Errors or unknown results may consume allowance: never automatically retry. No model is required. source_checked is source checking, not independent verification. Setup and validation limits

The steps

  1. Run manually
  2. Choose Skill
  3. Connection before
  4. Read budget
  5. Can read one detail?
  6. Skill detail
  7. Inspect evidence
  8. Connection after
  9. Evidence and balance receipt
  10. Read not started

Similar automations