fix(sandbox): redact temporary session credentials in model output

[icecrasher321]

Summary

• Carry host-issued callback credential provenance into the existing model-output registry when Function execution completes, including export failures. Scan returned JSON in place without an extra response copy.
• Keep these lease-scoped credentials out of durable file lineage so ordinary binary exports remain usable. Reuse the existing matching rules, placeholders, and provenance budgets.

Type of Change

• Bug fix

Testing

• 25 process acceptance checks with real Redis, covering JavaScript, shell, one-shot fallback, credential revocation, binary exports, export errors, nested results, deep traversal, and large ordinary results. Each output guard was removed independently and its regression checks failed; JSON reports retained locally.
• Focused application and session tests, application type-check, lint, all 52 audits, base-aware block registry check, and docs manifest check.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing (new tests pass the test-audit authoring gate)
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
docs	Skipped		Oct 1, 2026 2:18am UTC

[cubic-dev-ai]

All reported issues were addressed across 5 files

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[greptile-apps]

Confidence Score: 5/5

[Critical risk] Adds credential redaction to sandbox session output handling.

The PR appears safe to merge; no outstanding findings remain.

Summary

The PR registers lease-scoped sandbox callback credentials for model-output redaction, including export-error responses, without adding those credentials to durable file provenance. Since the previous review, it replaces the recursive output scan with an iterative one and adds deep-nesting tests.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Sandbox session issues callback credential] --> B[Sandbox execution returns JSON]
  B --> C[Scan output for credential]
  C --> D[Import matching provenance into model-output registry]
  D --> E[Compact and return response]
  C --> F[Keep credential out of durable file lineage]

Loading

_{Reviews (3) · Last reviewed commit: "fix(sandbox): scan session output withou..."}

[icecrasher321]

@greptile

[icecrasher321]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@icecrasher321 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

All reported issues were addressed across 5 files

<sub>Reply with feedback, questions, or to request a fix.

Fix all with cubic | Re-trigger cubic</sub>

[icecrasher321]

@greptile

[icecrasher321]

@cubic-dev-ai review this PR

[cubic-dev-ai]

@cubic-dev-ai review this PR

@icecrasher321 I have started the AI code review. It will take a few minutes to complete.

[cubic-dev-ai]

No issues found across 5 files

Confidence score: 5/5

• Automated review surfaced no issues in the provided summaries.
• No files require special attention.

_{Re-trigger cubic}