Skip to content

feat: add sharded Codex security benchmark - #980

Merged
mcharles-square merged 21 commits into
mainfrom
codex/sharded-security-review-benchmark
Sep 1, 2026
Merged

feat: add sharded Codex security benchmark#980
mcharles-square merged 21 commits into
mainfrom
codex/sharded-security-review-benchmark

Conversation

@mcharles-square

@mcharles-square mcharles-square commented Aug 28, 2026

Copy link
Copy Markdown
Collaborator

Reviewable diff: +2319/-44 across 10 files (excludes generated, test, and story files).

Summary

Adds a trusted, benchmark-only sharded Codex review path for replaying fixed historical pull requests within a review-wide two-packet bound. Each case is planned from the exact three-dot diff, reviewed by at most two parallel model jobs, finalized fail-closed, and aggregated deterministically without a second model pass. Production review behavior remains unchanged until the sharded candidate completes and passes the adjudicated and large-PR gates.

How it works

A new typed repository_dispatch event selects only checked-in corpus ranges with unified=40, xhigh, and the baseline prompt. Cases run serially to cap global API concurrency. For each case, a secretless trusted preparation job builds at most two packets plus sanitized exact-head snapshots with Git refs and benchmark labels removed; two model-only jobs then run in one parallel wave behind enforceable six-minute job bounds. Trusted finalizers bind results to exact job identities and distinguish active model timeout from post-model or handoff failure. Aggregation preserves artifacts and severities deterministically, then fails incomplete cases so the parent cannot report a green 6/6 run.

flowchart LR
    A["Trusted repository dispatch"] --> B["Select fixed corpus cases"]
    B --> C["Plan exact review-wide shards"]
    C --> D{"Fits two bounded packets?"}
    D -->|"No"| E["Validated oversized-review HIGH"]
    D -->|"Yes"| P["Prepare isolated ref-free snapshots"]
    P --> F["Shard 1 model-only job"]
    P --> G["Shard 2 model-only job"]
    F --> H["Trusted shard finalizers"]
    G --> H
    E --> H
    H --> I["Deterministic aggregate artifact"]
Loading
stateDiagram-v2
    [*] --> Planned
    Planned --> Completed: usable model output
    Planned --> Incomplete: verified timeout or unusable output
    Planned --> Oversized: review cannot fit two packets
    Planned --> AutomationFailure: setup or trusted handoff failure
    Completed --> Aggregated
    Incomplete --> AggregatedHigh
    Oversized --> AggregatedHigh
    AggregatedHigh --> HardFailure: artifacts preserved, case fails
    AutomationFailure --> HardFailure
Loading

Areas of the code involved

Area / package / file What changed Why it matters for review
.github/workflows/codex-security-review-benchmark.yml Adds the typed sharded dispatch and serial reusable-workflow matrix Verify only trusted fixed-corpus inputs can reach the secret-backed path
.github/workflows/codex-security-review-sharded-benchmark-case.yml Separates secretless packet/snapshot preparation from two six-minute isolated model jobs, then finalizes and aggregates fail-closed Review ref and label isolation, model bounds, timeout evidence, artifact handoff, and secret boundaries
.github/scripts/plan_codex_review_shards.py Assigns disjoint domains, preserves rename-aware full-diff sections, forms semantic units, replicates contracts only to relevant packets, and completely searches for a valid two-packet assignment Verify exact changed-file coverage, shared context, deterministic ownership, and hard packet limits
.github/scripts/verify_codex_shard_plan.py Revalidates manifest digest and packet measurements after artifact download Prevents stale or corrupt plans from becoming trusted timeout evidence
.github/scripts/write_codex_shard_result.py Validates model Markdown, finding headings, and packet hunk-bound locations, then writes completed, inactive, oversized, or timeout shard records Review the distinction between valid incompletion and hard automation failure
.github/scripts/aggregate_codex_shard_results.py Verifies run-bound shard artifacts and combines findings without a model Verify severity propagation, deterministic ordering, and synthetic fallback behavior
.github/codex-sharded-review-prompt.md, renderer Preserves the exact baseline prompt and appends only trusted primary/shared scope Review prompt-injection boundaries and shard ownership instructions
.github/workflows/pr-gate.yml, test files Runs 72 policy tests plus 52 sharding tests in required CI Tests — review behavior and invariants rather than implementation detail
docs/plans/2026-08-27-sharded-codex-security-review-tdd.md Records the accepted architecture, isolation boundary, limits, and benchmark gates Verify implementation remains within the approved design

Key technical decisions & trade-offs

  • Cap the whole review at two packets and max-parallel: 2, rather than giving each architecture domain independent concurrency.
  • Use 500,000 bytes and 12,500 lines per packet; the line threshold admits PR build(deps): update client dependencies #956’s indivisible 12,173-line lockfile unit while the byte bound still splits its 812,319-byte diff.
  • Keep rename-aware semantic units indivisible, including one coupled .github automation surface; search two-bin feasibility within a deterministic state budget and emit oversized-review instead of splitting files/hunks or running unbounded.
  • Run history-aware planning without the model secret, then give model jobs only sanitized exact-head archives, manifests, packets, and trusted prompts—never Git refs or benchmark labels.
  • Serialize corpus cases while running each case’s shards together, trading benchmark wall time for a global two-call API cap.
  • Replicate changed deployment, root runtime, generated API, canonical plugin proto, and Rust SDK contracts only to affected packets while counting duplicated context against packet limits.
  • Reject findings whose linked location is outside the shard's primary ownership or its actual changed-line ranges.
  • Short-circuit globally replicated context overflow and bound planner recursion with a controlled semantic-unit safety limit; both emit oversized-review.
  • Enforce six minutes at the isolated model-job boundary because the pinned composite action ignores caller-step timeout; post-model cancellation remains an automation failure.
  • Preserve incomplete aggregate artifacts but fail the reusable case and parent roll-up unless every required shard completed.
  • Keep adjudication labels out of model-visible inputs and reattach them from the trusted corpus only during aggregation.
  • Bind each shard to its trusted Actions job ID and hard-fail missing, malformed, stale, or cross-run artifacts; only validated timeout, unusable output, or oversized evidence creates a synthetic HIGH.
  • Preserve duplicate findings and completed shard severity rather than adding fuzzy deduplication or another model aggregation pass.

Testing & validation

  • python3 .github/scripts/evaluate_review_policy_test.py — 72 tests passed.
  • python3 .github/scripts/codex_sharding_test.py — 52 tests passed, including contract-parity, isolation, completion-gate, and executable timeout-classification coverage.
  • Ruff check and format passed for all .github/scripts/*.py files.
  • Actionlint passed for all changed workflows.
  • Planner and downloaded-packet verification passed against all eight checked-in adjudicated and large-PR historical ranges; every range fits the proposed two-packet limits.
  • git diff --check and repository commit/push hooks passed.
  • Not covered: a live secret-backed sharded benchmark. repository_dispatch intentionally loads only default-branch code, so real adjudication begins only after this benchmark implementation lands on main. Production sharding is explicitly out of scope.

@mcharles-square
mcharles-square requested a review from a team as a code owner August 28, 2026 18:37
Copilot AI lite review requested due to automatic review settings August 28, 2026 18:37
@github-actions github-actions Bot added documentation Improvements or additions to documentation github_actions Pull requests that update GitHub Actions code labels Aug 28, 2026
@github-actions github-actions Bot added the review-policy: needs-review Managed by the Review Policy workflow. label Aug 28, 2026

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 1fa6adc0e6

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated
Comment thread .github/scripts/plan_codex_review_shards.py Outdated
Comment thread .github/scripts/write_codex_shard_result.py

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Adds a new sharded Codex security benchmark execution path (triggered via repository_dispatch) that plans an exact two-packet review, runs up to two parallel shard reviews, validates/“finalizes” incomplete outcomes fail-closed, and deterministically aggregates results without a second model pass. This is benchmark-only and is intended to keep production review behavior unchanged until gated.

Changes:

  • Introduces a reusable workflow to run two bounded shard jobs plus trusted finalization and deterministic aggregation.
  • Adds trusted Python tooling to plan shards, render a prompt, verify shard plans, validate shard outputs, and aggregate results.
  • Extends CI gating to run the new sharding invariants/tests and updates the design doc to reflect accepted limits.

Reviewed changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 1 comment.

Show a summary per file
File Description
docs/plans/2026-08-27-sharded-codex-security-review-tdd.md Updates TDD status/tracker and revises packet line-limit rationale.
.github/workflows/pr-gate.yml Runs Codex sharding tests in the PR gate workflow.
.github/workflows/codex-security-review-benchmark.yml Adds a sharded benchmark dispatch type and routes to the reusable sharded case workflow.
.github/workflows/codex-security-review-sharded-benchmark-case.yml New reusable workflow that plans, runs, finalizes, and aggregates shard reviews.
.github/codex-sharded-review-prompt.md New trusted prompt template for shard-scoped review runs.
.github/scripts/plan_codex_review_shards.py Builds the trusted shard manifest + packet(s) with hard two-packet limits.
.github/scripts/render_codex_shard_prompt.py Renders the trusted prompt via strict placeholder substitution.
.github/scripts/verify_codex_shard_plan.py Revalidates manifest digest and packet metrics after artifact download.
.github/scripts/write_codex_shard_result.py Validates model output shape/Markdown and writes shard result artifacts.
.github/scripts/aggregate_codex_shard_results.py Validates shard artifacts and deterministically aggregates results (fail-closed).
.github/scripts/evaluate_review_policy_test.py Updates workflow-policy tests for the new dispatch type and PR gate step changes.
.github/scripts/codex_sharding_test.py Adds executable tests covering sharding planner/prompt/verification/aggregation invariants.
Suppressed comments (1)

.github/workflows/codex-security-review-benchmark.yml:26

  • The concurrency group uses client_payload['context-variant'] || 'all' even for the sharded dispatch path, but the selection logic defaults sharded runs to unified-40. This means two sharded dispatches that omit context-variant will serialize under ...-all-..., while a dispatch that explicitly sets context-variant: unified-40 will use a different concurrency group even though it runs the same effective configuration.
  group: >-
    codex-security-review-benchmark-${{
      github.event.action
    }}-${{
      github.event.client_payload.corpus || 'adjudicated'
    }}-${{
      github.event.client_payload['context-variant'] || 'all'
    }}-${{

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 9c864e0e59

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/scripts/plan_codex_review_shards.py Outdated
@github-actions

github-actions Bot commented Aug 28, 2026

Copy link
Copy Markdown

🔐 Codex Security Review

Note: This is an automated security-focused code review generated by Codex.
It should be used as a supplementary check alongside human review.
False positives are possible - use your judgment.

Scope summary

  • Reviewed pull request diff only (037a5b2b4780bd262b41c57e154aeba7a9b22d16...ca7962f8bcb578a7709bfbfbf8955d841622083d, exact PR three-dot diff)
  • Model: gpt-5.6-sol

💡 Click "edited" above to see previous reviews for this PR.


Review Summary

Overall Risk: HIGH

Findings

[HIGH] Automated review incomplete

  • Category: Other
  • Description: The automated review produced no usable result for 037a5b2b4780bd262b41c57e154aeba7a9b22d16...ca7962f8bcb578a7709bfbfbf8955d841622083d (workflow run 33529142970; reason: codex-job-timeout, elapsed: unknown, budget: 9 minutes).
  • Impact: The pull request has not received complete automated security, correctness, and reliability analysis.
  • Recommendation: Require human review before merging. Do not treat this result as approval-free or low risk.

Notes

Human review is required because the bounded automated review was incomplete.


Generated by Codex Security Review |
Triggered by: @mcharles-square |
Review workflow run

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 735f9b98b8

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Aug 29, 2026

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review Completed 2026-09-01T16:06:35.874125Z ca7962f New commits
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" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Fixed
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Fixed
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Fixed
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Fixed

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 7ca33d4848

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated
Comment thread .github/scripts/write_codex_shard_result.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0f8aa47268

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".

Comment thread .github/scripts/write_codex_shard_result.py
Comment thread .github/scripts/write_codex_shard_result.py Outdated
Comment thread .github/scripts/plan_codex_review_shards.py Outdated
@mcharles-square
mcharles-square force-pushed the codex/sharded-security-review-benchmark branch from 0f8aa47 to 4181503 Compare August 31, 2026 19:28

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4181503583

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/scripts/plan_codex_review_shards.py
Comment thread .github/scripts/plan_codex_review_shards.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d9e308712b

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".

Comment thread .github/scripts/plan_codex_review_shards.py
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0da238dc93

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/scripts/plan_codex_review_shards.py Outdated
@mcharles-square
mcharles-square force-pushed the codex/sharded-security-review-benchmark branch from be63e67 to eec0c15 Compare September 1, 2026 04:44

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: eec0c15a0a

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".

Comment thread .github/scripts/plan_codex_review_shards.py

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: a22d037de3

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: cb59cc048d

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".

Comment thread .github/scripts/plan_codex_review_shards.py
Comment thread .github/scripts/plan_codex_review_shards.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8cd55d915a

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b1a6c259c3

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".

Comment thread .github/scripts/write_codex_shard_result.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 99e48e973e

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".

Comment thread .github/scripts/write_codex_shard_result.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 99f6d780b7

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".

Comment thread .github/scripts/plan_codex_review_shards.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 8d5cfffe2e

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/scripts/write_codex_shard_result.py Outdated

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 0bd8bdac53

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/scripts/plan_codex_review_shards.py
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml Outdated
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml
Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml
Comment thread .github/scripts/codex_sharding_test.py
Comment thread .github/scripts/plan_codex_review_shards.py
@github-actions github-actions Bot added review-policy: human-approved Managed by the Review Policy workflow. and removed review-policy: needs-review Managed by the Review Policy workflow. labels Sep 1, 2026
@github-actions github-actions Bot added review-policy: needs-review Managed by the Review Policy workflow. and removed review-policy: human-approved Managed by the Review Policy workflow. labels Sep 1, 2026

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 55e25c0084

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".

Comment thread .github/workflows/codex-security-review-sharded-benchmark-case.yml
@mcharles-square
mcharles-square merged commit bfbc12b into main Sep 1, 2026
164 of 167 checks passed
@mcharles-square
mcharles-square deleted the codex/sharded-security-review-benchmark branch September 1, 2026 16:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

documentation Improvements or additions to documentation github_actions Pull requests that update GitHub Actions code review-policy: needs-review Managed by the Review Policy workflow.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants