Wispr Flow: Dictation That Works EverywhereStop typing. Start speaking. 4x faster.
Promoted
Trusty Squire
Sign up for APIs & websites within codex and claude
12 followers
Sign up for APIs & websites within codex and claude
12 followers
Trusty Squire lets Claude Code, Codex, Cursor, and other coding agents sign up for websites, sign in, configure services, and save API keys without putting them in chat or code.
Would love to see a way to scope which sites each agent can access, like team-level permissions or an allowlist per project. Right now it seems all-or-nothing, which makes it risky to let an agent loose on a bunch of accounts at once.
@feyzar7vk good feedback - yes enterprise level granular permissions haven't been implemented yet. The initial focus has been on solo operators and developers because it leads to a simpler MVP. But absolutely agreed team level access grants, budgets, permissions with granular admin controls and visibility is absolutely needed for more serious use
Would love to see a visual audit log for each agent action, like a little timeline showing when a key was used, by which agent, and on which site. That way I can quickly spot anything weird without digging through every conversation.
@zahide899987 the audit log tracks every creation, access, rotation etc (only screenshotted non sensitive stuff) but it could obviously be a lot more granular. Think this could evolve to be a very good control substrate for agents accessing credentials! (Where: trustysquire.ai/vault)
A credential rotation reminder would be huge, something that nudges you when API keys or passwords stored through Trusty Squire are getting old or might have leaked. Helps keep things tidy without you having to track it manually.
@efeqgl7 yeah actually this is pretty easy, will include in the next release. Thanks for the feedback!
the fact that you separated api keys from the actual chat context is honestly such a smart move, feels like it should have been a default this whole time
@muhammed592898 Thanks! The signup is only the first piece of it though. It can truly automate truly complex tasks -I've seen it clear and refresh the entire product line of a dropshipping store by hooking up to backends like cj dropshipping and dsers, and pulling in the best products by reviews.
Would love to see a shared vault feature so multiple agents on the same project can access the same saved credentials without each one needing to re-register separately.
the scoped secret handling feels like the right call, keeping API keys out of chat context is the kind of guardrail coding agents have been missing.
@cakelanur97361 exactly right. Password managers need to evolve to the agentic workflow where most credentials are not being directly created / consumed by the user. This is interesting because trusty squire limits the blast radius of what agents can do with credentials. They can't leak it or upload it to model companies servers.