Should every agent share one key?
No. Separate keys make usage attribution, revocation, and incident response cleaner.
Agent-native market data
Use product-level Cornerstones keys for agents and keep private upstream credentials in server-side runtime config. That makes revocation, usage tracking, and plan enforcement visible.
How should teams manage market-data access keys when coding and trading agents run repeatedly?
Separate upstream secrets from agent credentials; agents receive revocable product access while the runtime keeps private integrations locked down.
export CORNERSTONES_API_KEY='ck_...'
cornerstones-client auth login --api-key "$CORNERSTONES_API_KEY"
cornerstones-client verify| Option | Best for | Tradeoff |
|---|---|---|
| Cornerstones | Agent-native market context | Requires adopting a product-level access model |
| Raw private integration | Human-operated internal systems | Can expose credentials, adapters, and unstable implementation details to agents |
| Static prompt paste | One-off prototypes | No freshness, usage accounting, entitlement boundary, or repeatable evidence trail |
No. Separate keys make usage attribution, revocation, and incident response cleaner.
Use dashboard usage and quota surfaces rather than relying on local scripts alone.
Only if the account plan permits those routes. Free, Pro, and Max boundaries remain explicit.
Revoke or rotate the Cornerstones key, then inspect usage. Private upstream credentials stay protected.