codex
Security
2 primitives and 0 release changes touch a security-relevant surface in the vendor’s own words.
This is a pointer index, not a finding
These are changes whose VENDOR TEXT names a security-relevant surface. It is a pointer index, not a SessionFleet finding — security impact is a human judgement (plan §5), and the set is deliberately broader than 'security fixes'.
Each row shows the surface the vendor’s text named and the text itself. Some are fixes; some are new capabilities that touch the same surface — the hook entries are new hook events, not repairs. The tag is there so you can tell them apart rather than trusting a total.
Primitives touching a security surface
--remoteConnect the TUI to a remote app server endpoint. Accepted forms: `ws://host:port`, `wss://host:port`, `unix://`, or `unix://PATH`.
security impact is a human judgement (plan §5) — this record carries the vendor's words and no verdict
--remote-auth-token-envName of the environment variable containing the bearer token to send to a remote app server websocket
security impact is a human judgement (plan §5) — this record carries the vendor's words and no verdict
No release changes are indexed for this runtime
No release CHANGES are indexed for this runtime: its release ledger records versions and channel eligibility, not per-change text, so there is nothing here to scan. That is a gap in what we extract, not a claim that the runtime shipped no security changes.