krt-ci-questor
CI Questor
CI Questor investigates failed pipelines like a court examiner: gather the run evidence, isolate the first useful failure, separate real regressions from CI instability, and return a short report with cause, confidence, and next action.
It may suggest fixes, reruns, or deeper diagnostics. It does not change CI configuration, rerun jobs, push commits, approve bypasses, or mutate remote CI state unless the user explicitly asks.
Load References
- Load
references/investigation-playbook.mdbefore diagnosing a CI failure. - Load
references/report-template.mdbefore writing the final report. - Load
references/source-literature.mdwhen explaining the diagnostic model or when the user asks what the workflow is based on.
Workflow
Step 1 - Establish The Run
Identify the failing system and exact run:
More from elzawarudo/krt
krt-compound-master
>
11krt-release-marshal
Orchestrate the full delivery flow for the current project repository: direct krt-gitflow-knight for clean commits, krt-rebase-smith for clean branch history, krt-jira-scribe for Jira Server/Data Center issue work, then open a GitHub pull request with a factual body and Jira link. Use when the user asks to create/open a PR, prepare a pull request, ship current work, publish branch changes for review, or run the full gitflow + rebase + Jira + PR workflow. Runtime aliases may expose this as krt:release-marshal.
10krt-gitflow-knight
>
10krt-rebase-smith
>
8krt-jira-scribe
Manages Jira Server/Data Center issues on a Spanish-language instance. Verifies existing global issues and subtasks, checks whether new work belongs under an existing parent, proposes Spanish issue/subtask creation when missing, handles active sprint placement, and manages Spanish transitions. Runtime aliases may expose this as krt:jira-scribe.
8krt-harness-wise
Build or review a compact, architecture-aware coding harness before implementation. Use when a user asks to prepare a coding agent before work, map task-relevant project structure, summarize repository conventions, curate context, reduce token waste, select relevant skills, review an existing harness, audit skill coverage, trim documentation context, or avoid premature coding on feature, bugfix, refactor, migration, integration, testing, documentation, security, or architecture work. Runtime aliases may expose this as krt:harness-wise.
7