Unify Guardian context section collection (#42076)

## What changed

- Add typed context sections for root-conversation evidence, trusted user answers, and conversation transcripts to `codex-guardian-context`.
- Use the shared section registry when building core Guardian prompts and asynchronous Guardian V2 scoring context.
- Preserve source-role labeling while delivering authorization evidence as user-role context, and fail closed when context collection fails.

## Testing

- Cover shared section ordering and role preservation for synchronous and asynchronous targets.
- Verify trusted user answers appear exactly once as user-role input in Guardian V2 approval flows.

GitOrigin-RevId: 1d5d793ffbf166077411a66e85e039596907fb7e
This commit is contained in:
felixxia-oai
2026-09-01 15:22:21 +00:00
committed by copyberry
parent d141438342
commit 2350823caa
15 changed files with 357 additions and 108 deletions
+1 -24
View File
@@ -151,30 +151,7 @@ pub struct CodexThreadSettingsOverrides {
pub personality: Option<Personality>,
}
/// One root conversation message exposed only to a worker's Guardian reviewers.
#[derive(Debug, Eq, PartialEq)]
pub enum GuardianRootMessage {
/// Genuine root-user input that can establish or revoke authorization.
User(String),
/// Root assistant final output that provides untrusted conversational context.
Assistant(String),
/// Bounded, already role-labeled genuine user answers and their assistant questions.
UserInput(String),
}
impl GuardianRootMessage {
/// Renders every line with its original role so message content cannot impersonate another role.
pub fn render(self) -> String {
let (role, text) = match self {
Self::User(text) => ("user", text),
Self::Assistant(text) => ("assistant", text),
Self::UserInput(fragment) => return fragment,
};
text.lines()
.map(|line| format!("{role}: {line}\n"))
.collect()
}
}
pub use codex_guardian_context::GuardianRootMessage;
/// Authorization state that changes on genuine user input or history resets.
#[derive(Clone, Copy, Debug, Eq, PartialEq)]