Index of every page on this site: llms.txt
All skills
openai avatar

/security-scan

@ec0d6c1
by openaiopenai/codex-security14k stars
1,108
Run it yourself

Find vulnerabilities in the current source of a whole repository or selected paths. This is the default single-pass audit. Use the diff skill for PR, commit, branch, or working-tree changes and the deep skill for an explicitly requested deep scan.

Use this Skill: https://skilld.dev/gh/openai/codex-security/security-scan

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

referencessdk-files.md

≈548 tokens on demand. Your agent reads this file only when SKILL.md points to it.

SDK scans without a bound draft tool

Write unsealed scan-manifest.json and findings.json in the exact CODEX_SECURITY_SCAN_DIR. Put review bookkeeping in scan.reviewState. Save the model in an early draft even if findings are empty. Update these files at subsequent checkpoints and at the final handoff. Use the scan contract for unsealed fields. Each finding needs a stable identity.anchor. The SDK derives final IDs and fingerprints. Use the completed-scan examples for shape only.

Use the registered CODEX_SECURITY_SCAN_ID and the exact SDK values for scan.target: CODEX_SECURITY_TARGET_ID, CODEX_SECURITY_TARGET_DISPLAY_NAME, CODEX_SECURITY_TARGET_KIND, and the supplied revision and snapshot digest (CODEX_SECURITY_TARGET_REVISION, CODEX_SECURITY_TARGET_SNAPSHOT_DIGEST). Omit the digest for git_revision. Set the producer name to codex-security-plugin. Omit scan.sealedAt and scan.artifacts.

For a scoped target set scan.mode to scoped_path. After writing the manifest, run the supplied scope-binding command. If the SDK supplies only the registered environment, bind the exact scope file with the command for your shell.

"$CODEX_SECURITY_PLUGIN_ROOT/scripts/launch_codex_security_mcp" --helper bind-repo-scopes --scopes-file "$CODEX_SECURITY_TARGET_PATHS_FILE" --manifest "$CODEX_SECURITY_SCAN_DIR/scan-manifest.json"

In PowerShell, resolve the registered paths from the current location before calling CMD.

$env:scopePluginRoot = Convert-Path -LiteralPath $env:CODEX_SECURITY_PLUGIN_ROOT -ErrorAction Stop
$env:scopeScanDir = Convert-Path -LiteralPath $env:CODEX_SECURITY_SCAN_DIR -ErrorAction Stop
$env:scopeTargetPaths = Convert-Path -LiteralPath $env:CODEX_SECURITY_TARGET_PATHS_FILE -ErrorAction Stop
cmd.exe /d /v:off /s /c '""%scopePluginRoot%\scripts\launch_codex_security_mcp.cmd" --helper bind-repo-scopes --scopes-file "%scopeTargetPaths%" --manifest "%scopeScanDir%\scan-manifest.json""'

For a whole repository set scan.mode to repository.

Verify both files exist and return control to the SDK. It generates the report and threat model and seals the scan.

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts8d3 checks · Risk SAFE
  • Gen Agent Trust Hub8d

    This skill facilitates automated security audits by utilizing local scripts to analyze repository source code. It includes built-in safeguards, such as explicit instructions to treat input files as data and to avoid following external links, which help maintain a secure analysis environment. The use of platform-specific commands is consistent with the skill's role as a cross-platform development tool.

  • Socket9d

    No alerts

  • Snyk8d

    Risk: LOW · No issues

Signed by skilld at ec0d6c1. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated last week

README badge

README badge for openai/codex-security/security-scan