docs: define the trusted-local Codex Security threat model - #61
Conversation
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 01e3cdb4da
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ea18cb11a4
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Summary
Codex Security runs with your local permissions. This PR explains the practical security model: scan only repositories you trust and either own or are authorized to assess.
It also makes the scanner's behavior clear:
codex_security_scanfilesystem profile andapprovalPolicy: "never".--codexandcodexOverridesdo not make that profile or approval policy more restrictive.Why
Permission to assess a repository is not the same as trusting it. Configuration settings are not security guarantees when the scan does not apply them. Users should be able to see which actions and data flows are authorized before starting a scan.
Scope
The PR changes only:
SECURITY.mdsdk/typescript/README.mdIt does not change skills, scanner behavior, or runtime configuration.
Validation