About rhdh-overlay-cve-export
rhdh-overlay-cve-export is a skill for producing the CSV expected by RHDH CVE Management from overlay Git history. It directs users to a script for package-level CVE extraction and handles Jira enrichment separately.
Key features
- 01Overlay Git history extraction
- 02Plugin package and CVE pairing
- 03CSV export through a script
- 04Separate Jira data enrichment
- 05Output and completion checks
Use cases
- 01Preparing an RHDH CVE inventory
- 02Identifying plugin containers with CVE fixes
- 03Applying Jira fields to exported rows
Install
Copy the command and run it in a terminal.
npx skills add https:/ / github.com/ redhat-developer/ rhdh-skills --skill rhdh-overlay-cve-exportAfter installing, type/rhdh-overlay-cve-exportin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
agents/openai.yamlreferences/overlay-cve-csv-format.mdoverlay-cve-sources.mdscripts/compute-plugin-package-overlay-cve-list.mjstests/compute-plugin-package-overlay-cve-list.test.mjsworkflows/overlay-cve-export.mdDetails
From the project READMEPlugin overlay CVE export
Turn overlays git history into the CSV that RHDH CVE Management expects: one row per CVE and affected plugin package, scoped to generally-available workspaces.
This is a package-inventory job, not release reporting. It answers "which plugin
containers carry which CVE fix", from git — not "how is the release doing", which
is /rhdh-release-status.
Route
Load workflows/overlay-cve-export.md. It covers the extract run, Jira
enrichment, and the flag set.
Two references belong to this workflow alone and are read on demand:
references/overlay-cve-csv-format.md— the exact header, the row rule, the Container column, and the default output path.references/overlay-cve-sources.md— which workspaces and commits count, how CVEs and Jira keys are extracted, and how--sinceresolves.
Do not reimplement the parsing
scripts/compute-plugin-package-overlay-cve-list.mjs owns commit selection, CVE
extraction, pairing, dedupe, and CSV shape. Run it. Reading the overlays repo by
hand and assembling rows in chat produces a different answer than the export the
CVE process is checked against.
The script never calls Jira. Enrichment is a separate pass: invoke /rhdh-jira-api
for the issue fields, write them to a JSON file, and feed that back through
--apply-enrich.
Writing rules
The extract itself only clones and reads. Writing the enrichment JSON and the final CSV creates files on the user's disk — say where they will land and get agreement before the run that writes them. Never commit the CSV into a repository, and never post it anywhere.
Completion
Complete when the CSV path is named, the row count is stated, and the --since
revision the script actually resolved is reported rather than the version string
that was asked for. Every CVE without a Jira issue is listed with the overlays
commit URL that stands in for it. CVSS appears only where a ProdSec CVSS field
supplied it — a blank CVSS column is the correct output, never a filled-in guess.
Rows dropped by the --jql filter are counted, so the user can tell an empty
result from a filtered one. No credential or OAuth secret appears in any output.