mcpField
R
rhdh-jira-refine

rhdh-jira-refine

github.com/redhat-developerProductivity1952 installsUpdated 1d ago
Repository

About rhdh-jira-refine

rhdh-jira-refine audits existing RHDH Jira work against the workflow’s readiness checks. It is for teams reviewing issues and feature exploration, with read-only reports that distinguish missing data from fields that could not be retrieved.

Key features

  1. 01Per-issue readiness audits
  2. 02Feature Exploration checklist
  3. 03Read-only reports by default
  4. 04Reports JQL, issue count, truncation, and findings
  5. 05Enriches custom fields before evaluating completeness

Use cases

  1. 01Check whether Jira issues meet refinement criteria
  2. 02Find missing readiness information across issues
  3. 03Review Feature Exploration against workflow checks

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/redhat-developer/rhdh-skills --skill rhdh-jira-refine

After installing, type/rhdh-jira-refinein Claude Code to confirm it loaded.

Bundled filesagents/openai.yamlworkflows/refine-issues.md

Details

From the project README

Refine RHDH Jira work

Answer "is this ready" with the criteria the RHDH workflow actually enforces, and say what is missing rather than guessing at it.

Route

Load workflows/refine-issues.md. It covers both the per-issue readiness audit and the Feature Exploration checklist — the same checks, applied at different points in a Feature's life.

Not gated on grilling

Unlike /rhdh-jira-create, this skill does not require the external grilling skill and must not be blocked on it. Refinement reads work that already exists. Gating it on an optional external skill would take a whole capability offline for no benefit.

Reads by default, writes only on request

The report is read-only. When the user asks for fixes, each change becomes an external write: invoke /mutation-gate and follow it. A refine pass typically proposes many small fixes at once, so state them as one set — issue key and exact command per row — rather than approving them one at a time.

Boundary with the neighbouring skills

  • Opening new work is /rhdh-jira-create.
  • Updating one known key — status, comment, assignee, link — is /rhdh-jira-update. Refinement reports across many issues; it does not replace that.
  • Building the next sprint from the refined backlog is /rhdh-jira-sprint-plan.
  • Pre-refinement ceremony brief with smart analysis is /rhdh-prepare-refinement.
  • Exit criteria tables, field IDs, JQL, and the component catalog are /rhdh-jira-api.
  • Sizing scales, duplicate detection, and decomposition rules are /rhdh-jira-authoring.
  • What is still open against a release is /rhdh-release-status.

Completion

Complete when every issue the report covers was actually fetched with its custom fields enriched — an unenriched search returns empty Story Points, Team, Size, and Sprint, which looks identical to a field nobody set and is the single most common source of a false "missing data" finding. Report the JQL, the issue count, and whether the result was truncated. Every finding names the check that produced it and its severity. A field that could not be retrieved is reported as unretrieved, never as missing.

Comments

Markdown supported