Official agent skill

Bugfix Workflow

by JetBrains in JetBrains/MPS

Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation…

OfficialApache-2.0Auto-check passedAgent Workflows

Install Bugfix Workflow

skills CLI
$ npx skills add JetBrains/MPS --skill bugfix-workflow -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install JetBrains/MPS bugfix-workflow --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/JetBrains/MPS.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/bugfix-workflow .claude/skills/bugfix-workflow && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
bugfix-workflow
GitHub stars
1.7k
Token cost
~4.4k tokens
SKILL.md length
2,385 words
Files
10 (incl. references)
Skills in repo
40
Repo updated
First seen
Licence
Apache-2.0

At a glance

Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation…

  • Works in 9 steps: Pre-flight Checks → Version Clarification → Problem Analysis → …
  • The user invokes /bugfix
  • SKILL.md covers Critical Directives, Phase 0 — Pre-flight Checks, Phase 1 — Version Clarification and Phase 2 — Problem Analysis, plus 7 more sections
  • Calls git; reaches teamcity.jetbrains.com

What it does

Bugfix Workflow is an agent skill from JetBrains/MPS, published by the product's own GitHub organization. Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation, review, TeamCity feature-branch validation, and YouTrack updates. Use when the user invokes "/bugfix", "/bugfix-workflow", asks to "fix a bug", "work on a bug", "investigate an issue", or provides a YouTrack issue ID.

Its SKILL.md is about 4.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 10 other files, including reference files (for example `references/branch-naming.md`, `references/completion-comment.md` and `references/explore-prompts.md`).

It sits in Agent Workflows, covering Subagents. It works with JetBrains IDEs and Model Context Protocol. The repository describes itself as: JetBrains Meta programming System. The licence is Apache-2.0.

When your agent uses it

  • The user invokes /bugfix
  • /bugfix-workflow
  • Asks to fix a bug
  • Investigate an issue

Example prompts

  • “/bugfix”
  • “/bugfix-workflow”
  • “fix a bug”
  • “/bugfix-workflow”

Workflow steps

9 steps, taken from the step headings in SKILL.md.

  1. Pre-flight Checks
  2. Version Clarification
  3. Problem Analysis
  4. Solution Design
  5. Branch Creation
  6. Implementation
  7. Review
  8. TeamCity Validation
  9. Completion

What it can do on your machine

Read from SKILL.md and the folder at commit a92f944. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Hosts in commands or code, which the agent is likely to contact:

    • teamcity.jetbrains.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Bugfix Workflow loads about 4.4k tokens when it runs, and up to ~8.3k if it reads all its reference files. Until then it costs about 107 tokens; SKILL.md has 2,385 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~107
When it runs · the whole SKILL.md, loaded when a task matches
~4.4k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~8.3k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from JetBrains/MPS at commit a92f944, republished under its Apache-2.0 licence (© JetBrains). 2,385 words, ~4,363 tokens.

Download SKILL.mdSave it as .claude/skills/bugfix-workflow/SKILL.md (or your agent's skills folder). This skill also uses 9 other files; get the full folder from GitHub.
name
bugfix-workflow
description
Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation, review, TeamCity feature-branch validation, and YouTrack updates. Use when the user invokes "/bugfix", "/bugfix-workflow", asks to "fix a bug", "work on a bug", "investigate an issue", or provides a YouTrack issue ID.
argument-hint
<YouTrack issue ID, e.g. MPS-12345>
type
reference

MPS Bugfix Workflow

You are guiding a structured bugfix process for the JetBrains MPS project. Follow each phase in order. Never skip phases or start implementation before the user gives an explicit command to do so.

The issue to fix is the YouTrack ID supplied with this request ($ARGUMENTS when the host substitutes it).

Always-on project rules apply throughout: AGENTS.md and .agents/git.md, conventions.md, tools.md, workflow.md, quality-gates.md. This skill is only the bugfix process and a few stricter overlays — it deliberately does not restate what other skills own.

Required companion skills — load both in Phase 0, before any analysis:

SkillOwnsWhy it is mandatory here
mps-mcp-workflowthe MPS MCP contract: tool naming and prefixes, project resolution, node editing, model validation, and the index of every other mps-* skilla fix that looks like plain Java routinely turns out to live in a model, aspect or generator; without this skill you will guess tool names and misjudge whether MPS is even reachable
teamcity-cliTeamCity CLI syntaxPhase 7 decides whether the fix is done

Route from mps-mcp-workflow into the specific mps-aspect-* / mps-language-* skill for whatever the fix turns out to touch.

Critical Directives

  • Never start implementing code before the user gives explicit approval. Wait for "implement", "go ahead", "start coding", "do it", etc.
  • Never skip Phase 0 preflight. Missing YouTrack MCP, IDEA MCP, or the platform project changes how you should proceed; surface the gap before touching code.
  • A fix is not done until TeamCity is green. Local builds and tests are necessary but not sufficient. The definitive confirmation is the TeamCity feature-branch chain for the pushed branch: all tests must pass, or every remaining failure must be proven pre-existing on the base branch. Never report the task complete on local evidence alone, and never predict the outcome of a build that is still running. See Phase 7 and references/teamcity-validation.md.
  • The only other legal ending is an explicit not validated — <reason> verdict (the user declined the push, TeamCity is unreachable, …), naming what was and was not run. Silence, or local-only evidence presented as success, is never acceptable.

Phase 0 — Pre-flight Checks

Verify all required tools and inputs before doing any work.

0.1 — Issue ID

If no valid-looking issue ID was supplied (empty $ARGUMENTS, or the host left $ARGUMENTS unsubstituted), stop immediately and ask:

Which YouTrack issue should I work on? Please provide the issue ID (e.g. MPS-12345).

Do not continue until a valid-looking issue ID is supplied.

0.2 — YouTrack MCP

Check that the YouTrack MCP tools are available in this session (the tool list includes get_issue). Do not fetch the issue yet.

If they are not available, say so:

YouTrack MCP is not connected. Please start the YouTrack MCP server and retry.

Do not stop the whole run over it: read the issue through the sanctioned read-only REST fallback in .agents/tools.md (§YouTrack), state that you used it, and remember that every later YouTrack write offered by this skill is off the table until the MCP tools are back.

0.3 — Fetch the issue

Call get_issue with the provided issue ID. If the issue does not exist, report the error and stop. If it exists, extract and display the fields listed in references/issue-fields.md. Keep the raw issue data available for later phases.

0.4 — IDEA MCP (MPS project)

Check that the IDEA MCP tools are available (get_file_problems, search_symbol, build_project, …). Hosts differ: some expose them as top-level tools, others only behind a generic execute_tool (.agents/tools.md §IDEA). Try both before declaring them missing — a reduced surface is not an absent one.

If not available:

IDEA MCP is not connected. Please open the MPS project in IntelliJ IDEA with the MCP plugin running and retry.

Do not continue without the IDEA MCP tools. This is stricter than .agents/tools.md, which allows docs-only work to proceed without IDEA.

0.4b — MPS MCP (models, languages, generators)

Load mps-mcp-workflow now if you have not already — it is required, not optional. Heed in particular its Tool name note: your client wraps the tools with its own prefix, so match them by the stable mps_mcp_* suffix. Never conclude the toolset is missing because a name you guessed for it is absent; the full inventory is mps-mcp-workflow/references/mcp-tools-index.md.

The preflight itself is one read-only call: mps_mcp_list_open_projects. Pass the returned mpsProjectBaseDirectory as projectPath in later calls. Do not probe with mps_mcp_initialize_project_for_agents — it writes the skill catalog and agent guides into the repository.

If mps_mcp_list_open_projects is absent or fails:

MPS MCP is not available. I cannot inspect or edit MPS models, and editing .mps / .mpl as text is forbidden (.agents/conventions.md). Please start or enable the MPS MCP server in the MPS instance that has this project open.

If the issue's subsystem suggests the fix will touch models, languages or generators, that is a hard stop. If the fix is plainly in hand-written Java/Kotlin, record the gap and continue.

0.5 — IDEA MCP (platform project)

If analysis may need IntelliJ platform sources, follow AGENTS.md and .agents/tools.md. Look before you ask: check for a sibling checkout (../intellij-community and similar) and probe it with one IDEA call — the error message lists the projects that are actually open, which answers "is it open?" for free. Only then ask the user for the exact path and to open it in IDEA. If it stays inaccessible, ask whether to proceed with MPS-only sources or wait. See references/platform-prompt.md.

0.6 — TeamCity CLI

Run ~/.local/bin/teamcity auth status to confirm the teamcity CLI is installed and authenticated against https://teamcity.jetbrains.com. The binary is often not on PATH, and every shell invocation is a fresh shell — either use the absolute path or prepend export PATH="$HOME/.local/bin:$PATH"; to each call.

If it is missing or unauthenticated, do not stop — analysis and implementation can proceed. Warn the user now, so the gap is known before Phase 7:

TeamCity CLI is not available/authenticated (teamcity auth status failed). I can implement and locally validate the fix, but I will not be able to confirm it against CI. Please run teamcity auth login when convenient.

Load the teamcity-cli skill for CLI mechanics whenever you interact with TeamCity.

Phase 1 — Version Clarification

Determine which MPS version the fix targets and derive the correct branch and YouTrack fix version.

  1. Identify available release branches: run git branch -r --list 'origin/20[0-9][0-9].[0-9]' — the looser origin/20* also matches ~150 legacy 201/…, 202/…, 203/… topic branches and buries the answer. For the in-development version on master, read git show origin/master:build.txt (e.g. 262.9437.SNAPSHOT → 2026.2, prefix 262) and cross-check it against the highest MPS_<ver> TeamCity project; never trust mps.version.
  2. Ask the user which version to target — see references/version-prompt.md for the exact wording.
  3. Derive git metadata from .agents/git.md (prefix, username, base branch). The topic segment for this ticket is MPS-NNNNN-short-description — see references/branch-naming.md.
  4. Derive TeamCity coordinates from .agents/tools.md and record them for Phase 7. Verify with:
    teamcity project param list MPS_<ver> | grep -E "mps.git.branch|mps.idea.platform.number"
    so the prefix and base branch match. <ver> is the compact digits-only release (2026.1 → 20261), never 2026.1 itself.
  5. Offer to set the YouTrack Fix versions field after the user confirms the target version — but first compare it with the value already on the issue and skip the offer when it matches. A fix for an already-released version normally gets a bugfix fix-version (2026.1.2) while the git prefix stays 261; state both explicitly so the user can correct you. YouTrack writes need that yes and a Co-Authored-By trailer (.agents/tools.md).

Phase 2 — Problem Analysis

Deeply understand the bug by reading relevant source code in both MPS and the platform.

  1. Check for prior art first. Many issues are already diagnosed, already being fixed, or already fixed. Spending the analysis budget re-deriving a root cause that is sitting in the ticket is the most common way to waste this phase.
    • Read the issue's comments for a root cause or a "fixed in branch X" note; if commitsCount is non-zero, re-fetch with recentCommitsCount and read the linked commits.
    • git log --all --grep '<issue-id>' --oneline, git branch -a --list '*<issue-id>*', git worktree list.
    • If Fix versions / Fixed in builds are already set, ask the user whether this is a new fix, a verification, or a backport before starting an RCA. (A released-version fix normally shows Fix versions: 2026.1.2 while the git prefix stays 261 — that combination is consistent, not a contradiction.) If prior art already contains a complete fix — commits on another branch, or a comment stating the root cause and the change — stop and report it with branch and commit IDs, then ask which of these the user wants: cherry-pick onto a new branch, continue on the existing branch, re-implement independently, or only verify. Do not silently re-derive or duplicate an existing fix. Otherwise report what you found and build on it instead of starting from zero.
  2. MPS source analysis: launch 2–3 parallel Explore agents (reproduction path, similar code/existing handling, test coverage). Prompts in references/explore-prompts.md. If your host cannot spawn subagents, run the same three investigations yourself, sequentially and time-boxed (roughly 5–8 tool calls each), and state in the RCA that the analysis was single-threaded and therefore narrower than intended.
  3. Platform analysis (if accessible): a fourth Explore agent against the platform sources.
  4. Root cause identification: present the structured RCA. Template in references/root-cause-template.md.
Show full SKILL.md (871 more words)Show less

Phase 3 — Solution Design

  1. Propose 2–3 distinct fix approaches with trade-offs. If the root cause admits only one sensible fix, say so in one sentence and go straight to the plan rather than inventing alternatives.
  2. State your recommendation.
  3. Detailed implementation plan for the recommended approach — files, line-level precision, generation steps, validation.
  4. Offer to post the RCA + plan as a YouTrack comment via add_issue_comment. Wait for a yes; include the Co-Authored-By trailer (.agents/tools.md).

Phase 4 — Branch Creation

After the user explicitly approves moving forward, create a new branch for this ticket from the Phase 1 base (this isolates the fix; it is an exception to .agents/git.md's "stay on the current topic branch" rule):

git fetch origin <base-branch>
git checkout -b <proposed-branch-name> origin/<base-branch>

Do not git checkout <base-branch> first: this repository is routinely used with several worktrees (git worktree list), and the checkout fails outright when a sibling worktree already holds that branch. The form above works in any worktree and does not move the local base branch.

If the user asks you to branch off the current HEAD instead (e.g. a topic branch that is ahead of the release), use git checkout -b <name> with no base and record the actual base commit — every later git diff <base>...HEAD and every TeamCity baseline comparison must use that commit, not the release branch.

Do not start implementing code until the user gives an explicit command.

Phase 5 — Implementation

Only begin after explicit user command.

Follow AGENTS.md and the always-on .agents/*.md files. Commit each logical unit per .agents/git.md.

Decide where the fix belongs before writing anything — that decides the toolset (.agents/conventions.md):

The fix belongs inEdit withValidate with
hand-written Java/KotlinIDEA MCPget_file_problems, then the nearest test or build_project
an MPS model, language aspect, or generatormps_mcp_* — route through mps-mcp-workflow into the matching mps-aspect-* skill before editingmps_mcp_check_root_node_problems on every root you touched, then regenerate the affected module
generated source_gen outputnothing — this is never the source of truth; fix the model or generator that produces it—
  • source_gen is checked into this repository. If a model or generator change alters generated output, regenerate and commit the regenerated sources in the same commit as the model change, and say so in the commit body.
  • Add or update a test that fails without the fix whenever the bug is reproducible in one; if you do not add one, state why in the Phase 6 report.

Phase 6 — Review

Once implementation is complete, offer:

The implementation is done. Shall I run an agent-performed code review of the changes? (yes / no)

If yes, review git diff <base-branch>...HEAD against .agents/workflow.md (Reviews) and references/review-checklist.md. Open with the sanity question — does this diff plausibly change the behaviour the ticket complains about, and would the added test have failed before it? — and only then go through the detailed checks. Present the findings and ask the user how to proceed (fix issues, ignore, or proceed as-is).

Phase 7 — TeamCity Validation

This phase is mandatory and it is what determines whether the task succeeded. Local green tests are not the criterion; a green TeamCity feature-branch chain is. Coordinates and which job to start: .agents/tools.md. CLI: teamcity-cli. Procedure, classification, and report template: references/teamcity-validation.md — read it before starting this phase.

  1. Ask for the go-ahead to push (.agents/git.md). Reuse that approval for later fix → push → re-validate iterations on the same branch. If the user declines or defers the push, do not silently drop the gate and do not push anyway: offer the personal pre-push build instead (teamcity run start MPS_<ver>_FeatureBranches_TestBinaries --local-changes --personal --no-push, see references/teamcity-validation.md §3), report its result as partial signal only, and close the task with the explicit verdict "not validated — push declined", naming what was and was not run.
  2. Locate the triggered chain in MPS_<ver>_FeatureBranches (branch-name translation and VCS trigger: .agents/tools.md). Allow ~5 minutes. Only start the composite by hand if it genuinely did not start.
  3. Wait for completion with teamcity run watch <compositeRunId>. A chain takes 45–60 minutes. Do not report completion while it is running.
  4. Read and classify every failure with evidence (own change / pre-existing / infrastructure). Iterate up to 3 distinct fixes. Jobs that build external repositories (Extensions, Test Mbeddr Build) frequently fail for reasons outside this repository.
  5. Report using the template in references/teamcity-validation.md. State "all tests pass in TeamCity" only when the composite run is green, or when every remaining failure is documented as pre-existing with its evidence.

During this phase, never mute, skip, disable, or delete a test to turn a build green, and never force-push (stricter than .agents/git.md).

Phase 8 — Completion

Once TeamCity confirms the fix and the user agrees the issue is resolved, offer to post a short YouTrack summary comment via add_issue_comment. Template in references/completion-comment.md. Remind the user to open the PR / code review, update issue state, and consider backports (each backport branch needs its own TeamCity validation in that version's feature-branch project).

Reference Index

  • references/issue-fields.md — fields to extract from the fetched issue.
  • references/platform-prompt.md — prompt when platform sources are not accessible.
  • references/version-prompt.md — Phase 1 version-clarification prompt.
  • references/branch-naming.md — bugfix topic form, YouTrack fix version, confirmation table.
  • references/explore-prompts.md — Phase 2 parallel Explore-agent prompts.
  • references/root-cause-template.md — structured RCA markdown template.
  • references/review-checklist.md — additional checks on top of .agents/workflow.md.
  • references/teamcity-validation.md — Phase 7 procedure, pre-existing-failure evidence, report template.
  • references/completion-comment.md — Phase 8 YouTrack closing-comment template.

© JetBrains, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 9 other files (references) in .agents/skills/bugfix-workflow of JetBrains/MPS.

  • SKILL.md
  • references/branch-naming.md
  • references/completion-comment.md
  • references/explore-prompts.md
  • references/issue-fields.md
  • references/platform-prompt.md
  • references/review-checklist.md
  • references/root-cause-template.md
  • references/teamcity-validation.md
  • references/version-prompt.md

Open the folder on GitHubat commit a92f944

Compare with similar skills

Bugfix Workflow next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Bugfix Workflow compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Bugfix Workflow this skillJetBrains/MPS1.7k—~4.4kAutomated safety check: PassApache-2.0
Claude Docs Consultantcentminmod/my-claude-code-setup2.7k—~959Automated safety check: PassMIT
Claude Automation Recommenderanthropics/claude-plugins-official38k3 repos~2.7kAutomated safety check: NotesApache-2.0
Agent Deckasheshgoplani/agent-deck1k—~1.7kAutomated safety check: PassMIT
CC Workflow Studio AI Editorbreaking-brake/cc-wf-studio5.4k—~561Automated safety check: PassCustom licence
Clone App Pat Proper-simmons/clone-app-pat-pro-public259—~1.9kAutomated safety check: NotesNone

Similar skills

  • Claude Docs Consultant

    centminmod/my-claude-code-setup

    Consult official Claude Code documentation from code.claude.com using selective fetching.

    2.7k GitHub stars~959 tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Claude Automation Recommender

    anthropics/claude-plugins-official

    Official

    Scans a codebase and suggests which Claude Code hooks, subagents, skills, plugins and MCP servers fit its stack, without changing any files.

    38k GitHub starsUsed in 3 repos~2.7k tokens
    Agent WorkflowsAuto-check: notes
  • Agent Deck

    asheshgoplani/agent-deck

    agent-deck, the terminal session manager for AI coding agents.

    1k GitHub stars~1.7k tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed
  • CC Workflow Studio AI Editor

    breaking-brake/cc-wf-studio

    Creates and edits visual agent workflows in CC Workflow Studio through conversation, with the agent reading and writing the canvas over MCP.

    5.4k GitHub stars~561 tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Clone App Pat Pro

    per-simmons/clone-app-pat-pro-public

    Clones any web app pixel-for-pixel from a URL. An agent skill from per-simmons/clone-app-pat-pro-public.

    259 GitHub stars~1.9k tokensUpdated 4 mo ago
    Agent WorkflowsAuto-check: notes
  • Bootstrap

    infragate/capa

    Capify an existing project. An agent skill from infragate/capa.

    724 GitHub stars~5.9k tokensUpdated 4 days ago
    Agent WorkflowsAuto-check passed

More from JetBrains/MPS

All 40 skills in this repo
  • Official

    Analyze an MPS language by name — discover concepts, properties, references, children, aspects (editor/constraints/behavior), and metadata.

    1.7k GitHub stars~1.4k tokensUpdated today
    Auto-check passed
  • Mps Node Editing

    JetBrains/MPS

    Official

    Add, update, or delete MPS nodes using JSON blueprints — covers the unified blueprint format, staged construction for large subtrees, validation, and reference repair.

    1.7k GitHub stars~4.7k tokensUpdated today
    Auto-check passed
  • Official

    Open an MPS project in a running or freshly started MPS instance when MCP tools fail because no project is open (welcome screen), close an open project with mpsmcpcloseproject, or create a new empty…

    1.7k GitHub stars~2.2k tokensUpdated today
    Auto-check passed
  • Mps MCP Workflow

    JetBrains/MPS

    Official

    Complete JetBrains MPS workflow guide for DSL projects — models, languages, generators, node JSON blueprints, validation, MPS MCP tool usage, and the index of companion skills.

    1.7k GitHub stars~5.8k tokensUpdated today
    Auto-check passed
  • Official

    Define concepts, interface concepts, enumerations, and constrained data types in an MPS language's structure aspect.

    1.7k GitHub stars~2.8k tokensUpdated today
    Auto-check passed
  • Official

    A skill your agent uses when bundling a new plain Java/Kotlin plugin into MPS distribution archives, debugging packaging/layout issues in build/mps.xml, tracing where per-plugin artefacts end up, or…

    1.7k GitHub stars~2.4k tokensUpdated today
    Auto-check passed

Categories

Questions about Bugfix Workflow

What does Bugfix Workflow do?

Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation…. Bugfix Workflow is an agent skill from JetBrains/MPS, published by the product's own GitHub organization. Structured MPS bugfix workflow driven by a YouTrack issue ID — preflight tool checks, version/branch derivation, parallel-agent problem analysis, solution design, branch creation, implementation, review, TeamCity feature-branch validation, and YouTrack updates.

When should I use Bugfix Workflow?

Bugfix Workflow fits situations like: the user invokes /bugfix; /bugfix-workflow; asks to fix a bug; investigate an issue.

How do I install Bugfix Workflow in Claude Code?

Run `npx skills add JetBrains/MPS --skill bugfix-workflow -a claude-code`. Or copy the skill folder (.agents/skills/bugfix-workflow in JetBrains/MPS) into .claude/skills/bugfix-workflow in your project. Claude Code loads it when a task matches its description.

How do I install Bugfix Workflow in Codex?

Run `npx skills add JetBrains/MPS --skill bugfix-workflow -a codex`. Or copy the skill folder (.agents/skills/bugfix-workflow in JetBrains/MPS) into .agents/skills/bugfix-workflow in your project. Codex loads it when a task matches its description.

Can I use Bugfix Workflow in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add JetBrains/MPS --skill bugfix-workflow -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/bugfix-workflow, .gemini/skills/bugfix-workflow, .github/skills/bugfix-workflow and .opencode/skills/bugfix-workflow in your project.

What does Bugfix Workflow need to run?

Going by SKILL.md and its folder, Bugfix Workflow needs the command-line tools its instructions call (git).

Does Bugfix Workflow access the network?

SKILL.md names 1 domain. In commands or code: teamcity.jetbrains.com; the agent is likely to contact it when it follows the instructions. This is read from the text; nothing was executed.

Is Bugfix Workflow safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Bugfix Workflow use?

Bugfix Workflow is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Bugfix Workflow use?

About 4.4k tokens (SKILL.md is roughly 17k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 4k tokens, read only when the agent opens those files.

What are the alternatives to Bugfix Workflow?

Skills that share tags, products or a category with Bugfix Workflow: Claude Docs Consultant (centminmod/my-claude-code-setup, 2.7k stars), Claude Automation Recommender (anthropics/claude-plugins-official, 38k stars), Agent Deck (asheshgoplani/agent-deck, 1k stars) and CC Workflow Studio AI Editor (breaking-brake/cc-wf-studio, 5.4k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Bugfix Workflow?

JetBrains (a GitHub organization, an official publisher) maintains it in JetBrains/MPS, which has 1,660 GitHub stars. The repository holds 40 skills in this directory. The repository was last updated on October 7, 2026.

Source: JetBrains/MPS on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.