Hacker Newsnew | past | comments | ask | show | jobs | submit | new23d's commentslogin

Our report seeks to answer some of our questions for seven of the most popular agentic code editors and plugins. By intercepting and analysing their network flows across a set of standardised tasks, we aim to gain insight into the behaviour, privacy implications, and telemetry patterns of these tools in real-world scenarios. Incidentally, a side-effect was running into OWASP LLM07:2025 System Prompt Leakage for three of the chosen coding agents. You can see the system prompts in the appendix.


Obfuscation via egress firewalls and evasive binary development with an iterative LLM agent.


Use AWS Route53?


Making a dynamic DNS client with aws and jq CLI, with a least-privilege IAM role and a SystemD service.


Exactly the same happened with me. Picking up the phone and responding to email (in weeks, not hours or days) didn't lower my bills. This sort of marketing is perhaps deflection.


We'll be working on that in the coming days. Thought the data at this point was a good start.


Some initial observations:

• Google's CRLs from the same intermediate CA (same public key) have different URLs and different content when pulled from different hosts (google.com, youtube.com).

• DigiCert has sharded according to 'assurance' class, algorithm, year and acquisition's name.

• Sectigo also has sharded according to 'assurance' class [1].

• GlobalSign has sharded by the yearly quarter presumably.

• HTTP Cache-Control maxage (or s-maxage), 'Expires' and 'Next Update' within the CRL file are not in sync.

• Some CAs other than Let's Encrypt also do not publish CRL URLs in the leaf certificates.

[1] https://www.sectigo.com/knowledge-base/detail/Sectigo-Interm...


We collected some data [1] on the viability of only CRLs as the future (phasing out OCSP) - motivated by Let's Encrypt's announcement today [2].

Data is on CRL availability, number of entries, expiry & refresh times, etc. from various x509 leaf server SSL certificates.

[1] https://news.ycombinator.com/item?id=41058138 [2] https://news.ycombinator.com/item?id=41046956


We collected some data on the viability of only CRLs as the future (phasing out OCSP) - motivated by Let's Encrypt's announcement today [1].

Data is on CRL availability, number of entries, expiry & refresh times, etc. from various x509 leaf server SSL certificates.

[1] https://news.ycombinator.com/item?id=41046956


TLS 1.3 and ESNI (now called Encrypted Client Hello - ECH) are separate standards, although you'll see ECH only enabled in bleeding edge stacks. In fact, ECH is still in IETF draft phase [1].

It can be disabled if an organisation wishes to. I wrote about how to do this in Chrome [2,3], and will write about Firefox when I get a chance.

[1] https://datatracker.ietf.org/doc/draft-ietf-tls-esni/ [2] https://chasersystems.com/blog/disabling-encrypted-clienthel... [3] https://news.ycombinator.com/item?id=37823262


Ah, i didn't know they actually moved eSNI out of TLS1.3. it was certainly part of quite a few drafts.

I do find it sad it isn't pushed harder. Companies who need to do interception have legitimate concerns, but they can be addressed.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: