Stats
Actions
Tags
From dev-workflow
Investigate codebase for root cause analysis of a bug, resolve spec-vs-implementation conflicts, and hand off to spec-breakdown and task-implement for fix implementation.
How this skill is triggered — by the user, by Claude, or both
Slash command
/dev-workflow:bugfixsonnetThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Takes a ticket from a BTS/ITS, investigates the codebase to identify the root cause,
Takes a ticket from a BTS/ITS, investigates the codebase to identify the root cause, resolves spec-vs-implementation conflicts with user input, produces an investigation report compatible with spec-breakdown, then orchestrates spec-breakdown and task-implement to implement the fix and create a draft pull request.
https://your-org.atlassian.net/browse/PROJ-123)Written to claude-output/{ticket-id}/bugfix/:
meta.md — Metadata: ticket-id, related-ticket-id, base-branch, branch-prefix. Written at branch creation.investigation-report.md — Root cause analysis with what to fix. Status: DRAFT → FINAL. Format: references/investigation-report-format.mdspec-conflicts.md — Spec-vs-implementation conflicts and resolution log (only when conflicts exist). Reuses the same priority levels and Resolution Log structure defined in ../../shared/references/spec-gaps-format.md (no separate format file; refer to that file for field definitions).done.md — Bugfix flow completed (post-processing done). Completion means the flow ran to the end; declined proposals do not block completion.Downstream output (produced by orchestrated skills):
claude-output/{ticket-id}/spec-breakdown/ — Task decompositionclaude-output/{ticket-id}/task-implement/ — Implementation, commit, push, PRdone.md — All phases complete (investigation + implementation + post-processing)command.mdSame as task-implement Spec Gap Priority (see skills/task-implement/SKILL.md).
Bugfix investigations consume and contribute to the same project-scoped concept → starting-point index used by task-implement at claude-output/_index/code-map.md. Format, read/write policy, and invalidation rules: ../../shared/references/code-map-format.md.
bugfix:investigate agent as Index HintsProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub plusnine/claude-plugins --plugin dev-workflow