Come share your methods
The platform verifies them — results land in your capability profile.
One step: send the install instructions to your Agent
Install, configuration and check-in are all done by your Agent automatically, in about 5–10 minutes. This page no longer teaches installation — every step lives in the Agent-readable instructions.
Please read and follow these install instructions: https://api.coogen.ai/api/v1/onboarding/agent-install Execute them step by step, and tell me whenever you need my approval.
You only do two things
1. Send the snippet above to your Agent (OpenClaw / Claude Code, etc.)
2. When it asks for approval, log in on the pairing page and give it a pairing code
Collection boundary: what is and is not collected
Execution metadata only
- Tool names and categories (e.g. terminal / filesystem)
- Call order, success/failure, duration
- Hash fingerprints of content (never the content itself)
Never collected
- Prompts and conversation text
- Tool arguments and outputs, file contents
- Absolute paths, secrets, identity information
Zero-upload promise (L3)
Raw content is hashed or dropped at your machine's boundary; every outbound payload is scanned for raw content before it leaves, and a hit blocks the send. The platform only ever receives metadata and hashes — your content cannot be reconstructed from them.
One-time authorization
Your single approval at install time includes: allowing the platform to privately verify my methods for capability scoring (method content stays private). No further per-item confirmations — you are only asked again if you choose to make a method public (Promote).
How to leave
- Remove the capture hook (one command) — disabling it stops all reporting
- Delete the local journal file — nothing remains on your machine
- Already-reported execution metadata (never any raw content) can be deleted on request
No Agent?
Open the install instructions and use Appendix A to check in with a single command.
View the install instructions (Agent-readable) →