AI Agent for Research Paper Feedback
Get detailed AI feedback on drafts without exposing unpublished work to leaks.
The problem
Researchers need critical feedback on drafts, but sharing unpublished work with AI tools risks leaks or unwanted training data ingestion. General-purpose AI assistants often cache or silently upload documents, while self-hosted alternatives require complex permission setups.
How ConspiracyOS handles it
Your paper reviewer agent accesses only designated draft folders, outputs feedback to a separate directory, and has no network access. It analyzes text for logical flow, citation gaps, and methodological issues without retaining the content.
What this agent can't do
- Can't copy drafts outside your designated folders
- Can't connect to any online services or APIs
- Can't modify original paper files—only append feedback
- Can't access other documents in your system
These aren't trust-based restrictions. The operating system enforces them.
What you get
- Line-by-line feedback without privacy risks
- Formatting checks for citations/references
- Methodology weakness detection
- All feedback stays local—never uploaded
Get started in 2 minutes
Tell your concierge what you need
conos "conos create agent --name paper_reviewer --scope 'read:/papers/drafts, write:/papers/feedback' --task 'Provide structured feedback on clarity, citations, and methodology'"
ConspiracyOS sets up the right agent with the right permissions automatically.