Adversarial code review run by a second model on your own machine.
What this Grok bot does
Runs an adversarial review over code, plans or a review pipeline, with the reviewer living on your own Hermes agent instead of the model that wrote the work. Grok Bot packages the context and the test plan, the reviewer works without seeing the author's reasoning, and every finding has to come from a real run or a named line.
Hermes Adversary at a glance
Hermes Adversary is an on demand coding bot shared by @PixelRainbowNFT and first listed on 2026-10-17. It works when you message it rather than on a timer, which keeps usage low. One public catalog lists it, and its x.ai share link was last checked live on 2026-10-17.
Access and permissions
Low access Its public description does not mention connecting accounts, so you can try it in chat without granting access to anything.
No account access mentioned in the public description.
Derived from the public description. The live configuration on x.ai is the source of truth — review it before adding.
Before you add Hermes Adversary
- Give review bots read-only repo access; let humans merge.
- Open the x.ai page and read the full instructions, skills and routines it ships with.
Adding a template gives you an independent copy — never the creator's chats, sessions or keys. How adding a Grok Bot template works.
More from @PixelRainbowNFT
Commander's Intent
@PixelRainbowNFT
Writes one shared plan your whole agent fleet can read, then keeps it current.
Outbound Prospecting + X Scout
@PixelRainbowNFT
Sourced prospect lists, first drafts, and a twice-daily scan of X for buyers.
X Posting Curator
@PixelRainbowNFT
Runs each X post through sourcing, hooks in your voice, fact-checks and a check against X's ranking weights