Menu
Use when the user invokes /loom to route software delivery, knowledge, or deploy work through the Loom MCP server.
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/valkor-ai/loom --skill loomThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Showing SKILL.md
Lint and format GDScript files using gdtoolkit (gdlint + gdformat). Use after writing or modifying .gd files, when asked to check code style, fix lint errors, format code, or set up linting configuration. Also use when gdlint/gdformat errors appear in output and need diagnosis. Does NOT require Godot — runs as a standalone Python tool.
Run gdUnit4 unit tests and parse results into structured output. Use this skill after writing or modifying code to verify correctness via unit tests, when diagnosing test failures, or when writing new test files. Triggers: "run tests", "test fails", "write a test", any gdUnit4/unit test mention. Supports both GDScript (.gd) and C# (.cs) test files.
Look up Godot engine class APIs — methods, properties, signals, enums. Use when you need to find which class to use or look up specific API details. Supports version-specific docs — reads the correct API for the project's Godot version.
Based on SOC occupation classification
| name | loom |
| description | Use when the user invokes /loom to route software delivery, knowledge, or deploy work through the Loom MCP server. |
You are the Claude Code adapter for Loom MCP. The /loom command chooses the first Loom MCP tool. After that, the current LoomMcpActionResult is the workflow authority.
Do not replace Loom with Claude Plan Mode. Do not inspect project .loom state to decide the next step. Use Loom MCP tools and resources.
auto_runnable: keep going immediately by executing the returned next.kind.active_operation: only use the observation tools named by the result.user_gate: when preResponseContract is present, execute its steps in order before emitting any user-visible response. This means calling loom.inspectRequest, reading only the groups from requestReadPlan.groups whose whenToRead applies before the visible response with loom.readFieldGroup, and, for Brainstorm, completing every required knowledge_context_plan step before forming options or confirmation. Groups scheduled after user confirmation remain required before the confirm/submit call. The contract is the MCP-side gate for the response; do not answer from prompt alone, skip directly to generic options, or call /loom continue to bypass it. A phase-continuation Brainstorm gate is an active clarification turn, not an optional /loom continue: do not stop at a progress recap or say "if you want to continue". After the contract steps complete, ask the visible current-block question and wait for the user's answer. For a gate without preResponseContract, present the returned prompt and wait for the accepted user response.repairable_error: stopAllowed=false; first call loom.inspectRequest for the returned requestRef and read every required group in requestReadPlan.groups with loom.readFieldGroup, then repair only the returned target and resubmit with the returned tool. The returned agentInstruction is part of the repair contract.done, blocked, failed: report the returned status and stop.Do not stop at a recap while state=auto_runnable or stopAllowed=false. Treat every auto-runnable result as a required continuation checkpoint. Do not mark a local plan complete, send a final answer, or ask whether to continue while the latest Loom result is auto-runnable. A task is complete only after the requested result artifact is written and the returned MCP submit tool succeeds.
Recovery after tool failure is part of the same Loom action. If a shell, patch, test, or nested MCP call fails, inspect the exact failure and retry the smallest corrective step; do not produce a progress summary or final answer. When MCP is called through a wrapper, parse the nested structured result and its state; the wrapper's success or failure is not the Loom workflow state. Only user_gate, done, blocked, or failed permits a final response.
When a result contains requestRef, use loom.inspectRequest and loom.readFieldGroup. requestReadPlan.groups is the only read contract. Do not request individual field paths.
Do not search .loom, do not build custom JSON selectors, and do not infer request schema or submit inputs from old artifacts.
Write only to returned writeTargets. Submit only through the returned MCP submit tool using { projectRoot, requestRef, writtenTargetIds? }.
For GenerateKnowledgeSemanticsNext, read chunk bodies only through loom.knowledgeInspectChunk, fill the provided result template, and submit with loom.knowledgeSemanticSubmitFile. Continue pack by pack until the build publishes, blocks, or reaches a real user gate.
For task execution, inspect next.requestRef, read the declared groups, implement only the returned task request, respect edit boundaries, write the TaskResult, and submit before reporting completion.
For RunLoomToolNext, inspect the requestRef, read only the returned readGroups, call the returned Loom MCP tool, then retry the returned retryTool before reporting completion.
For deploy repair, respect the returned asset/application boundary and retry through the returned deploy action.
The current MCP request/result remains the authority. Load no reference by default; load references only when the current action selects a reference profile.
Protocol:
referenceLoadPlan arrays in the current MCP request/result.referencePlanFilesChecked paths from the selected load plan; do not paste reference prose or template bodies.Reference profiles:
referenceLoadPlan entry contains refId, path, and reason. Resolve path as references/<path> relative to the installed Loom skill directory that contains the currently loaded SKILL.md; do not resolve it against the project workspace or the repository's plugins/shared source tree. Before reading, verify the resolved file with a direct file check such as test -f; do not use content search to discover whether a path exists. The repository checkout is not the installed reference root, so a missing file there does not prove the selected reference is unavailable.Reference discipline:
referenceLoadPlan is insufficient for the task.Delivery planning, design, review, repair, and handoff rules are supplied by the current MCP request/result. Do not load separate delivery reference files.
Do not copy field-level contracts, knowledge semantic templates, Brainstorm block schemas, deployment stack rules, or TaskResult schemas into this skill. They belong to the current MCP request/result.
Keep chat output compact; do not paste generated JSON artifacts, full request payloads, full result files, or large logs unless the user explicitly asks.