dsh-deepbrain

DeepBrain stores judgments — conclusions distilled from meetings, interviews, and lectures — each backed by an evidence chain that traces back to the exact words spoken and corroborated across separate occasions. dsh-deepbrain brings that judgment layer into DeepSeek Harness so your agent can pull verified conclusions, attribute word-for-word quotes, and look up a person's historical positions without ever leaving the chat.

Rather than exposing all twenty-seven server-side tools, the plugin registers six carefully scoped ones so the model does not default to grabbing raw transcripts and blowing up its own context. Start with deepbrain_brief when you are unsure what to ask, rely on deepbrain_topics for most writing and analysis, use deepbrain_quotes for attributable lines, run deepbrain_person before a negotiation or meeting, and keep deepbrain_call as an escape hatch for anything the other five do not cover.

It suits teams and individuals who need sourced analysis over generated filler: strategy research, competitive analysis, negotiation prep, and academic writing where misattributing a quote to the wrong person is a real risk. DeepBrain's evidence chains and cross-occasion corroboration make every claim traceable back to what was actually said.