Feedback Triage Chrome extension security report

Feedback Triage is a thinking tool for academics processing received feedback — supervisor emails, journal peer review, committee notes, editor decision letters, grant reviewer comments. You paste the feedback you received, highlight the parts that matter, and the tool scaffolds the cognitive work of turning each one into a concrete action with success criteria. YOU DO THE DECODING. The tool gives you the structure. There is no AI in this extension. No LLM call, no auto-classification, no AI-generated response paragraphs, no "smart" suggestions. Every piece of judgement — what the reviewer is actually worried about, how severe the concern is, what specifically you will do — is yours. The product exists to develop the meta-skill of reading critique well, not to do it for you. EVERYTHING STAYS ON YOUR DEVICE. The extension does not read the pages you browse and runs no content scripts on any external site. Your pasted feedback, verbatim quotes, interpretations, action plans, and private notes all live in chrome.storage.local on your machine and are never transmitted. The only network call the extension can make is to GradSummit's sign-up service — and only if you explicitly enter an email and click Subscribe for product updates. In that case the request body is exactly { email, source: "feedback-triage" }; no feedback content is ever included. HOW IT WORKS 1. Create a triage session by picking the kind of feedback you received — supervisor email, journal peer review, conference review, thesis committee feedback, editor decision letter, grant reviewer comments, reading group, or other. 2. Paste the full text of the feedback into the session. 3. Highlight any portion of the source text and click "Make this a feedback item." The selection becomes a verbatim quote attached to a structured card. 4. For each item, fill the card: • Verbatim text (extracted automatically, editable) • Type (methodological, framing, citation, clarity, ...) • Severity — Critical / Important / Cosmetic, three tiers that actually distinguish • Underlying concern — what is this person actually worried about, beneath the surface complaint? • Specific action — what will you do, concretely? • Optional: success criteria, effort estimate, status, and a private internal note that NEVER appears in any export. 5. Export. Choose: • Plain action list (Markdown) — bare list of actions grouped by severity • Response to Reviewers letter (Markdown) — standard journal format with reviewer quotes, bracketed prompts where your prose goes, and a professional closing • Revision action plan memo (Markdown) — for supervisor / committee / editor / grant feedback • Plain text — for email pasting • PDF — for print or attachment All export formats are included. There is no Pro tier and no paywall. Every feature works for everyone, free. THE REFERENCE FRAMEWORK LIBRARY Five short essays, written for graduate researchers and shipped inside the extension: 1. How to read peer review like a reviewer — common reviewer phrases and the underlying concern beneath each 2. Supervisor signal-to-noise — decoding "interesting," "have you considered," and the question-as-directive 3. Severity calibration — how to tell critical from important from cosmetic, with worked examples per source type 4. Underlying concern excavation — four techniques for getting from surface complaint to real worry 5. Decided not to address — when declining feedback is legitimate, and the professional how Total ~4,500 words of curated content. Static, on-device, no AI. WHO IT'S FOR PhD students and postdocs processing journal/conference peer review, supervisor emails, or thesis committee feedback. Faculty processing peer review and grant reviewer comments. Anyone in a feedback-rich knowledge-work role who is tired of treating every piece of critique as one undifferentiated blob. WHO IT'S NOT FOR Undergraduates (lower volume of structured feedback). Anyone looking for an AI that will write their response letter — this tool will refuse, by design. WHAT YOU WON'T FIND HERE • No AI of any kind (severity suggestion, type classification, response-paragraph drafting, summarization — none of it) • No tracking, telemetry, analytics, A/B testing, or advertising identifiers • No account system, no login, no cloud sync (your data is local — back it up via Settings → Export) • No host permissions for any third-party website. The one host permission declared is for our own sign-up Worker, used only if you opt in to product updates. PRIVACY POSTURE Reviewer comments are often confidential under journal policy. Supervisor candor is not meant to leave the room. Your private thoughts about feedback are personal. We treat all of it that way: stored locally, never transmitted, never seen by us. The Internal Note field on each item is verified by automated test to be absent from every export format — see the Privacy Policy for the full guarantee. A GRADSUMMIT TOOL Built by Dr. Rafiq Muhammad, PhD. Part of a family of small, focused tools for graduate researchers. Each one does a single thing well; none of them collect your data. Learn more at www.gradsummit.com.