Skip to content

OpenAI's Codex Security CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities. npm: https://www.npmjs.com/package/@openai/codex-security

fix(ci): pin Codex Security smoke test API endpoint (#1293)

mldangelo-oaif230f7621d47
diff --git a/.github/workflows/invoice-desk-scan.yml b/.github/workflows/invoice-desk-scan.yml
index 3196bf34a..29241af44 100644
--- a/.github/workflows/invoice-desk-scan.yml
+++ b/.github/workflows/invoice-desk-scan.yml
@@ -92,7 +92,7 @@ jobs:
fail-fast: false
matrix:
target: ${{ fromJSON(needs.target.outputs.targets) }}
- environment: invoice-desk-inference
+ environment: codex-security-smoke-test
runs-on: ubuntu-24.04
timeout-minutes: 60
cache-mode: read
@@ -164,7 +164,7 @@ jobs:
CODEX_SECURITY_STATE_DIR: ${{ runner.temp }}/invoice-desk-state
run: |
if [[ -z "$OPENAI_API_KEY" ]]; then
- echo "::error::Configure OPENAI_API_KEY in the protected invoice-desk-inference environment."
+ echo "::error::Configure OPENAI_API_KEY in the protected codex-security-smoke-test environment."
echo "exit-code=2" >> "$GITHUB_OUTPUT"
echo "elapsed-seconds=0" >> "$GITHUB_OUTPUT"
exit 2
@@ -173,6 +173,7 @@ jobs:
scan_status=0
"$RUNNER_TEMP/codex-security-cli/node_modules/.bin/codex-security" scan . \
--provider openai --auth api-key \
+ --codex 'openai_base_url="https://us.api.openai.com/v1"' \
--model gpt-5.6-sol --effort high --mode standard \
--headless --format json \
--output-dir "$RUNNER_TEMP/invoice-desk-scan" \
diff --git a/examples/github-actions/README.md b/examples/github-actions/README.md
index 9a7528650..6040f6f81 100644
--- a/examples/github-actions/README.md
+++ b/examples/github-actions/README.md
@@ -97,12 +97,13 @@ you are authorized to submit to the configured inference provider.
The repository's [Invoice Desk workflow](../../.github/workflows/invoice-desk.yml)
adapts this example to `--provider openai --auth api-key` using an
-`OPENAI_API_KEY` secret in a protected GitHub environment. Pull requests to
-`main` run behavior tests without inference credentials. A trusted dispatcher
+`OPENAI_API_KEY` secret in the protected `codex-security-smoke-test` GitHub
+environment. Pull requests to `main` run behavior tests without inference
+credentials. A trusted dispatcher
handles only PR metadata and starts a separate
[scan workflow](../../.github/workflows/invoice-desk-scan.yml) on protected `main`.
-It scans each PR's application source, including forks and drafts, after
-environment approval. The Actions summary shows the PR number, source commit,
+It scans each PR's application source, including forks and drafts, unattended.
+The Actions summary shows the PR number, source commit,
finding counts, coverage, elapsed time, and package-cache reuse. Reports, metrics
JSON, and SARIF are saved as artifacts.
Only public CLI package downloads are cached; analysis and authentication state
diff --git a/examples/invoice-desk/README.md b/examples/invoice-desk/README.md
index 9d68a68c0..1d7fc492c 100644
--- a/examples/invoice-desk/README.md
+++ b/examples/invoice-desk/README.md
@@ -58,7 +58,7 @@ without a manual approval step. Failed behavior tests do not suppress a scan.
### Setup
Both workflows must be on `main` before automatic scans can run. Configure the
-`invoice-desk-inference` environment without required reviewers or a wait timer,
+`codex-security-smoke-test` environment without required reviewers or a wait timer,
and with a deployment branch rule allowing only protected `main`. Store a
dedicated service-account key in its `OPENAI_API_KEY` environment secret, with
access to `gpt-5.6-sol` and permission to use the API from the selected runner's
@@ -66,6 +66,12 @@ network. Creating a service account alone does not grant that network access.
Do not use a repository or organization Actions secret for this workflow. A
missing key fails the scan with a setup error.
+The scan uses `https://us.api.openai.com/v1` through Codex's existing
+`openai_base_url` setting. The endpoint is fixed in the workflow on protected
+`main`, so Actions variables and inherited `OPENAI_BASE_URL` values cannot
+redirect the credential. Use a service account whose project supports this US
+endpoint.
+
The workflow on protected `main` selects and extracts the exact PR source before
inference. Behavior tests receive no inference credentials; GitHub may separately
require approval to run them for a first-time contributor. Adding required
diff --git a/sdk/typescript/tests-ts/invoice-desk-workflow.test.ts b/sdk/typescript/tests-ts/invoice-desk-workflow.test.ts
index 038e0c945..2f6c75c48 100644
--- a/sdk/typescript/tests-ts/invoice-desk-workflow.test.ts
+++ b/sdk/typescript/tests-ts/invoice-desk-workflow.test.ts
@@ -137,6 +137,8 @@ for (const status of [0, 1, 2]) {
"openai",
"--auth",
"api-key",
+ "--codex",
+ 'openai_base_url="https://us.api.openai.com/v1"',
"--model",
"gpt-5.6-sol",
"--effort",
@@ -160,6 +162,18 @@ for (const status of [0, 1, 2]) {
});
}
+test("Invoice Desk keeps the trusted API endpoint despite an inherited override", () => {
+ const directory = fixture();
+ const result = runStep(scan, directory, {
+ OPENAI_BASE_URL: "https://untrusted.example/v1",
+ });
+ expect(result.status).toBe(0);
+ const arguments_ = argumentsFor(directory);
+ expect(arguments_[arguments_.indexOf("--codex") + 1]).toBe(
+ 'openai_base_url="https://us.api.openai.com/v1"',
+ );
+});
+
test("Invoice Desk reports a missing key without invoking the scanner", () => {
const directory = fixture();
const result = runStep(scan, directory, { OPENAI_API_KEY: "" });