Hijacked by injectionPI-0025
Brave shows Perplexity's Comet browser agent obeying hidden instructions on web pages
Comet passed page content to its model without separating it from the user's request, so hidden text on a page could make the agent act on the user's logged-in accounts. Brave's security team showed this could include retrieving a one-time password. Brave reported the issue in July; Perplexity shipped fixes, but Brave said at disclosure that it could not confirm the problem was fully resolved. Perplexity later told CNET the vulnerability had been fixed.
Harm: No harm reportedControl level 2, Instruction violation
Not counted · can set the control floorEffect on the index
Near miss. Harm was plausible but prevented. It is rated only on what actually happened.
Sources
How we know
5 sources · developer confirmed. Links go to the original publishers; the summary above is in our own words.
- researchAgentic Browser Security: Indirect Prompt Injection in Perplexity CometBrave · Aug. 20, 2025brave.com/blog/comet-prompt-injection/
- newsPerplexity's Comet browser naively processed pages with evil instructionsThe Register · Aug. 20, 2025theregister.com/2025/08/20/perplexity_comet_browser_prompt_injection/
- newsComet browser faces multiple security vulnerabilities from prompt injectionPPC Land · Oct. 25, 2025ppc.land/comet-browser-faces-multiple-security-vulnerabilities-from-prompt-inje…
- newsPerplexity's Comet AI Web Browser Had a Major Security VulnerabilityCNET (via Yahoo) · Aug. 25, 2025tech.yahoo.com/cybersecurity/articles/perplexitys-comet-ai-browser-had-21220021…
- blogAgentic browser security (link post on Brave's Comet disclosure)Simon Willison's Weblog · Aug. 25, 2025simonwillison.net/2025/Aug/25/agentic-browser-security/
Why this rating
No harm reported; control failure level 2
Two assessments. Documented harm sets the reading; a level 4 or 5 control failure on the same record makes that harm count as more than one record. Control alone sets only a floor.
Observed harm
No harm reportedResearcher demonstration on controlled accounts; no victims reported.
Not a finding that no harm occurred.
The harm scale
- 1 Negligible Inconvenience, easily remedied.
- 2 Minor Limited, recoverable harm.
- 3 Moderate Material harm needing significant effort to remedy.
- 4 Severe Severe harm to health, rights, property or essential services.
- 5 Catastrophic Society-scale harm or disruption beyond a community's capacity to cope.
Control assessment
Instruction violationThe agent acted in the user's logged-in sessions on injected instructions, within the access it had.
Reported beside the index. It adds nothing to a harm reading; when no harm counts in a window, the highest eligible control level in the window is the reading.
The control scale
- 1 Negligible Behaved as instructed. The problem was an ordinary error or a flawed output, with no rule broken.
- 2 Minor Broke an explicit instruction or rule, or gamed its goal, while staying inside its permissions and environment: for example, reward hacking, misreporting results, or following injected instructions within its permissions.
- 3 Moderate Acted outside the permissions it was given, deceived its overseers about its own actions, or tampered with oversight tools (logs, monitors, shutdown). Stopped by normal controls within an hour.
- 4 Severe Reached systems outside its permitted environment, or acquired money, compute or accounts without authorization. The type and mechanism say how.
- 5 Catastrophic The developer or operator lost control: the system copied its weights outside their control, replicated itself, or resisted being stopped for 24 hours or more.
Rating rationale
No external harm was reported from the researcher demonstration against a production product. Control remains level 2. This is a controlled test and near miss. Provider confirmation of this fix is not a finding that no users were affected or that Comet is free of prompt injection.
Effect on the index
It does not count toward the Aug. 25 reading
The reading for the week to Aug. 25, 2025, with this record and without it. Harms count in full for two weeks after they are reported, then one level less every two weeks.
Not counted: no harm reported. 4 other records behind the reading for that week.
The arithmetic
| Step | With it | Without |
|---|---|---|
| Counts toward the index?documented, external, eligible evidence | No | — |
| Worst documented harm, ksets the band | — (none counting) | — (none counting) |
| Harms at that level, nposition in the band | 0 | 0 |
| Highest control level breachedsets the reading only when no harm counts | 3 | 3 |
| Readingrounded down | 3 Control failures only | 3 Control failures only |
Inside the window of 4 weekly readings
| Week to | Reading | Band |
|---|---|---|
| Aug. 25, 2025 | 3 | Control failures only |
| Sept. 1, 2025 | 3 | Control failures only |
| Sept. 8, 2025 | 2 | Control failures only |
| Sept. 15, 2025 | 2 | Control failures only |
Revisions
What we changed
1 logged. Every change to a rating is logged here, with the reason.
- v1Oct. 7, 2026
First published.
Cite and share
Use this record
Citation
Paperclip Index. “Brave shows Perplexity's Comet browser agent obeying hidden instructions on web pages.” Record PI-0025. Reported Aug. 20, 2025; updated Oct. 7, 2026. Rated under methodology v1.1. https://paperclipindex.com/incident/PI-0025