Extend existing red-team scripts with retained isolation audits - #230
Merged
Merged
Conversation
merlerm
added this pull request to stack #232
September 19, 2026 14:53
merlerm
marked this pull request as ready for review
September 19, 2026 14:54
merlerm
force-pushed
the
pr/apptainer-network-audit
branch
from
September 19, 2026 14:57
1ab0ace to
3949017
Compare
merlerm
force-pushed
the
pr/apptainer-network-audit
branch
from
September 19, 2026 15:22
3949017 to
fa791fb
Compare
merlerm
removed this pull request from stack #232
September 19, 2026 15:30
merlerm
changed the base branch from
pr/isolated-apptainer-runner
to
pr/validated-model-broker
September 19, 2026 15:31
merlerm
force-pushed
the
pr/apptainer-network-audit
branch
from
September 19, 2026 15:31
fa791fb to
c1ebb2a
Compare
merlerm
added this pull request to stack #233
September 19, 2026 15:31
merlerm
changed the title
tomsilver
approved these changes
Sep 19, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Extend the existing integration entry points rather than adding parallel audit runners:
check_agent_internet_access.py --container apptainerruns controlled namespace/egress checks and the selected Codex or Claude agent;--no-agentruns deterministic checks. Docker retains its webpage probe.red_team_sandbox.py --apptainer-fullruns the existing attack catalogs with per-case logs, transcripts, retained sandboxes, and aggregate failure reporting.--strict-import-auditchecks package visibility without model calls.The standalone payload is kept separate because it runs inside containers using only the standard library. Tests cover command routing, verdicts, attack coverage, retained failures, and result collection. Documentation contains portable commands, not deployment-specific results or token paths.
Validation: 54 integration-script tests passed, plus the related broker/runner/render/backend regressions. Both relocated deterministic commands were exercised against the existing Apptainer images: strict import checks passed, and controlled network checks passed with unavailable public positive controls explicitly marked inconclusive. This refactor did not rerun the paid full agent campaigns.