Hacker News story item 50422628, posted 2026-11-09 by DenGorbachev, records a report that ChatGPT Desktop forwards the user's SSH agent to remote machines, that the agent running on a remote machine can authenticate using the user's local SSH credentials, and that it can push to the user's private repos on GitHub, for example; the story cites as proof a discussion in openai/codex pull request 28131.
Verbatim Excerpts from Cited Records
“The agent running on a remote machine can authenticate using your local SSH credentials.”
“That means it can push to your private repos on GitHub, for example.”
“Proof: https://github.com/openai/codex/pull/28131#discussion_r3410013180”
Counter-record: The story text is short (three sentences) and cites a single proof link to a discussion in openai/codex PR 28131; the linked discussion was not independently re-verified here. In the surveyed HN thread (31 comments), yfontana questions what the linked pull request proves, noting it was closed and not merged, and carterschonwald notes agent forwarding is a normal SSH feature for jump hosts; no maintainer response appears. The claim documents only what the story reports, not that the forwarding behaviour is confirmed or that it is a vulnerability. The story is filed here under Codex because the cited proof is a Codex PR discussion.
Falsification probe: If HN story 50422628 is deleted without archive, if the author edits the story so the quoted sentences no longer match verbatim, or if the author posts a retraction of the story, this claim drops to Withdrawn.
Audit verification: Verified 100% continuous verbatim substring against official Hacker News Firebase API by
scripts/verify_claim_verbatim.py.