Saltar para o conteúdo
Tecrübetecrube.com
First steps

Security & Approval

Tecrube's security model: OAuth 2.1, changes applied on your request and undoable, approval for suggestions, budget guardrail, audit log and results the assistant cannot invent.

Three principles

  1. The change you ask for is applied at once and can be undone. Your explicit request is the consent; the assistant does not ask again and changes only what you asked for. Every change is recorded, and changes that have a reverse can be taken back with undo_action. Permanent deletions are confirmed with you in chat first.
  2. The assistant's own suggestions wait for your approval. Changes you did not ask for (automation rules in "ask" mode, proactive suggestions) go to the approval list; the assistant first calls approve_action for a preview and applies it with confirm=true only after you explicitly approve that preview. The nightly scan does maintenance only (syncing, renewing authorisations, re-queueing stuck productions); it does not change budgets or publishing.
  3. The result tells the truth. Every write tool returns applied (with an undo record), not applied + reason (plan limit, budget guardrail, spend cap), pending approval (suggestions only), dry run (no connection) or failed + reason. The assistant must have seen this result before it says "applied".

Identity and authorisation

  • The assistant connection uses OAuth 2.1 + PKCE; there are no API keys. Clients register dynamically (DCR) and receive only the mcp scope.
  • Platform connections (Google, Meta, LinkedIn, TikTok, Microsoft) go through the provider's own OAuth screen; Tecrube never sees your password. The ChatGPT Ads key and the WordPress application password are stored encrypted in the panel.
  • Tokens are specific to a member and an assistant and are revoked individually from the panel. A call with a revoked token is rejected.
  • In team accounts every member connects with their own identity; the audit log records who called what.

Budget guardrail

Every budget change passes two limits; a request beyond them is not applied even if approved, and says why:

LimitRuleExample
Single stepA budget rises by at most 50% per change100 → at most 150
24 hoursThe total increase in one day cannot exceed twice the starting budgetA campaign starting the day at 100 cannot pass 200 that day

Undo is not caught by this limit: undo_action returns to the previous budget. In addition, the spend guard (set_spend_guard) lets you set daily/monthly caps, target CPA and margin; automation rules stay within those caps.

Audit log

Every tool call is recorded: which member, which assistant, which tool, parameters, duration, credits spent, result and error. Write operations additionally keep the preview, approval time and application result in the action log. Records are visible in the panel under Usage and Actions; the list_recent_actions tool brings the same log into chat.

Real undo

"Undo" does not just delete the record in Tecrube; a real reverse request goes to the platform: back to the previous budget, resume after pause, remove added negative keywords, unpublish, restore a WordPress revision, revert a GTM version. Operations with no reverse (a sent message, a removed campaign) return not_undoable with the reason; the tool says so instead of erroring.

Data and privacy

  • Tecrube passes tool results to the assistant; it does not store your chat content. The assistant provider's (Anthropic, OpenAI…) own data policy applies.
  • Platform tokens are stored encrypted and decrypted on the server only during a tool call.
  • Details: Privacy policy and data protection notice.

Next steps