Agent skill

Pure Backend Start

by finos in finos/legend-engine

Starts (or confirms) the standalone legend-engine backend - engine Server + H2 + local metadata server on fixed ports 9095/9092 - so a Pure LSP started with -Dlegend.test.

Apache-2.0Auto-check passed

Install Pure Backend Start

skills CLI
$ npx skills add finos/legend-engine --skill pure-backend-start -a claude-code

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

GitHub CLI
$ gh skill install finos/legend-engine pure-backend-start --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/finos/legend-engine.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/pure-dev/skills/pure-backend-start .claude/skills/pure-backend-start && 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
pure-backend-start
GitHub stars
113
Token cost
~1.2k tokens
SKILL.md length
529 words
Files
1
Skills in repo
15
Repo updated
First seen
Licence
Apache-2.0

At a glance

Starts (or confirms) the standalone legend-engine backend - engine Server + H2 + local metadata server on fixed ports 9095/9092 - so a Pure LSP started with -Dlegend.test.

  • The user wants to run Pure queries through the real engine
  • SKILL.md covers Why this exists, The launcher, Run the script and Then: point the LSP at it, plus 1 more section
  • Calls mvn
  • Through plan-gen

What it does

Pure Backend Start is an agent skill from finos/legend-engine. Starts (or confirms) the standalone legend-engine backend - engine Server + H2 + local metadata server on fixed ports 9095/9092 - so a Pure LSP started with -Dlegend.test. routes execute()/go() through real plan-generation instead of the interpreted engine. Starts only the backend, not the LSP. Use when the user wants to run Pure queries 'through the real engine', 'through plan-gen' or 'via the backend server', reproduce compiled-mode behaviour the interpreted engine cannot, or mentions legend.test.server…

Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

The licence is Apache-2.0.

When your agent uses it

  • The user wants to run Pure queries through the real engine
  • Through plan-gen
  • Via the backend server
  • Reproduce compiled-mode behaviour the interpreted engine cannot

Example prompts

  • “through the real engine”
  • “through plan-gen”
  • “via the backend server”
  • “/pure-backend-start”

What it can do on your machine

Read from SKILL.md and the folder at commit 43f4e69. 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:

    • mvn

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

  • Network

    No URLs in SKILL.md.

    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

Pure Backend Start loads about 1.2k tokens when it runs. Until then it costs about 142 tokens; SKILL.md has 529 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~142
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k

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 finos/legend-engine at commit 43f4e69, republished under its Apache-2.0 licence (© finos). 529 words, ~1,246 tokens.

Download SKILL.mdSave it as .claude/skills/pure-backend-start/SKILL.md (or your agent's skills folder).
name
pure-backend-start
description
Starts (or confirms) the standalone legend-engine backend - engine Server + H2 + local metadata server on fixed ports 9095/9092 - so a Pure LSP started with -Dlegend.test.* routes execute()/go() through real plan-generation instead of the interpreted engine. Starts only the backend, not the LSP. Use when the user wants to run Pure queries 'through the real engine', 'through plan-gen' or 'via the backend server', reproduce compiled-mode behaviour the interpreted engine cannot, or mentions legend.test.server, EngineServerForTest or LegendTest.

Start the standalone engine backend (H2 + metadata server + engine Server)

Why this exists

The Pure LSP go()/pure-lsp-go runs the interpreted engine, which (a) can't reproduce compiled-mode behaviour (plan generation + generated Specifics.prepare Java, e.g. relational graph-fetch temp-table nodes) and (b) has its own gaps (e.g. interpreted RFPM graph-fetch is unimplemented). meta::pure::router::execute(...) wraps the query in the mayExecuteLegendTest native which — if legend.test.server.host/port/clientVersion/serverVersion System properties are set on the JVM — does real plan-generation and POSTs to a running backend Server instead. This skill boots that backend on fixed ports; pure-lsp-launch-engine then points the LSP JVM at it via --jvm-arg=-D....

Result: go() through the so-configured LSP runs the same plan-gen + execute path as Test_*_UsingPureClientTestSuite, verified end-to-end (returns real query results and shows generated SQL in the go() output trace), without a 15-30 min mvn test.

The launcher

org.finos.legend.engine.ide.EngineServerForTest (in the legend-engine-pure-ide-light-http-server module, alongside PureIDELight). It starts H2 (fixed 9092), a local test metadata server (dynamic), and the full org.finos.legend.engine.server.Server on a FIXED port (9095), then blocks forever.

Run the script

All of check-if-already-up / compile-once / generate-classpath-once / launch-in-background / poll-for-ready is implemented deterministically in bin/pure-backend-start:

bash
pure-backend-start [--port N] [--h2-port N] [--force-classpath]

Defaults: --port 9095, --h2-port 9092. If something's already listening on --port, it reports that and exits 0 without relaunching (it's a full engine JVM — never stack a second one).

On success it prints one JSON line to stdout, e.g. {"port": 9095, "h2Port": 9092, "state": "ready"}, and exits 0. Exit codes:

  • 1 — the launcher module didn't compile, or its runtime classpath is missing/empty (pass --force-classpath to regenerate the cached classpath after a dependency change)
  • 3 — launched but never reached READY within the timeout (~2 minutes) — check /tmp/engine-server.log

Then: point the LSP at it

This skill only brings up the backend — it does not start the Pure LSP. pure-lsp-launch-engine calls pure-backend-start automatically by default and wires the resulting -Dlegend.test.server.* jvm-args into the LSP launch; pass --no-backend to skip this and launch interpreted-only instead. For a custom/manual launch, use pure-lsp-server's --jvm-arg passthrough with the flags pure-backend-start prints (see references/lsp-devloop-internals.md).

Show full SKILL.md (205 more words)Show less

Things worth knowing / gotchas

  • The go() wrapper must seed the DB and use the right execute FQN — createTablesAndFillDb() first, meta::pure::router::execute not meta::pure::mapping::execute, and one go() per session (reuse welcome.pure). All three are the database-setup and welcome.pure conventions in references/lsp-devloop-usage-rules.md.
  • Pure-only edits do NOT require rebuilding/restarting this backend Server — because plan generation runs LOCALLY in the LSP's interpreted JVM and only plan execution is delegated here. That is how legendExecute in core/pure/protocol/vX_X_X/invocations/execution.pure is wired (see pure-lsp-go's "Execution routing" section): let planGenerated = if(true, {|local executionPlan(...)}, {|remote generatePlan(...)}) keeps plan-gen local, so it uses the Pure graph in the LSP session (reflecting files you've pure-lsp checked live, on top of the jar baseline), and only executePlan($host,$port) hits this server. Plan execution is almost all Java, so it ignores your Pure edits. Net loop with NO mvn install / NO server restart: edit .pure → pure-lsp check it into the session → go() → live local plan-gen + server execute. (Rebuild/ restart this server only for Java execution-layer changes, or if you flip plan-gen to run remotely by editing that if(true,...).)
  • clientVersion vX_X_X is the dev placeholder and is accepted (see PureClientVersions).
  • Logs: /tmp/engine-server.log. Stop it with pkill -f EngineServerForTest (match the java process precisely; a bare pgrep can match your own shell command).

© finos, 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

Just SKILL.md in plugins/pure-dev/skills/pure-backend-start of finos/legend-engine.

Open the folder on GitHubat commit 43f4e69

Compare with similar skills

Pure Backend Start 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.

Pure Backend Start compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Pure Backend Start this skillfinos/legend-engine113—~1.2kAutomated safety check: PassApache-2.0
StartDonchitos/Claude-Code-Game-Studios26k—~6.5kAutomated safety check: PassMIT
Developing StandaloneTriliumNext/Trilium38k—~3.4kAutomated safety check: PassAGPL-3.0
Session Start Hookasgeirtj/system_prompts_leaks69k1 repos~1.2kAutomated safety check: PassCC0-1.0
Save As Standalone HTMLasgeirtj/system_prompts_leaks69k—~1.4kAutomated safety check: PassCC0-1.0
Where Do I Startmohitagw15856/pm-claude-skills1.4k—~912Automated safety check: PassMIT

Similar skills

  • Start

    Donchitos/Claude-Code-Game-Studios

    First-time onboarding — asks where you are, then guides you to the right workflow.

    26k GitHub stars~6.5k tokensUpdated yesterday
    Game DevelopmentAuto-check passed
  • Developing Standalone

    TriliumNext/Trilium

    A skill your agent uses when working on Trilium's standalone in-browser build (apps/standalone, deployed at app.triliumnotes.org and embedded by the mobile app) — the service worker (sw.ts), the…

    38k GitHub stars~3.4k tokensUpdated today
    DatabasesAuto-check passed
  • Session Start Hook

    asgeirtj/system_prompts_leaks

    Creating and developing startup hooks for Claude Code on the web.

    69k GitHub starsUsed in 1 repo~1.2k tokens
    DevelopmentAuto-check passed
  • Save As Standalone HTML

    asgeirtj/system_prompts_leaks

    Single self-contained file that works offline. An agent skill from asgeirtj/system_prompts_leaks.

    69k GitHub stars~1.4k tokensUpdated yesterday
    Frontend & DesignAuto-check passed
  • Where Do I Start

    mohitagw15856/pm-claude-skills

    Turn a chaotic pile of everything-in-your-head into one clear first action — the antidote to the paralysis of too much at once.

    1.4k GitHub stars~912 tokensUpdated yesterday
    Productivity & AutomationAuto-check passed
  • Getting Started

    live-codes/livecodes

    Quick start for standalone app at livecodes.io, embedding playgrounds with CDN or npm, and self-hosting basics.

    1.5k GitHub stars~1.6k tokensUpdated yesterday
    AI & LLM EngineeringAuto-check passed

More from finos/legend-engine

All 15 skills in this repo
  • Pure Chain Update

    finos/legend-engine

    Procedure for git-syncing and rebuilding a local legend-pure + legend-engine checkout, in dependency order, with pinned dependency versions resynced.

    113 GitHub stars~1.8k tokensUpdated today
    Auto-check passed
  • Pure Code Style

    finos/legend-engine

    Reformats a .pure file (or a block of Pure code) to a consistent house style: 2-space indentation with no column-aligned hanging indents, tight colons in type annotations, spaced pipes on…

    113 GitHub stars~1.6k tokensUpdated today
    Auto-check passed
  • Pure Lsp Check

    finos/legend-engine

    Compiles/checks a single .pure file against the already-running Legend Pure LSP bridge and reports diagnostics (errors/warnings) in under a second, as a fast alternative to a full mvn test/mvn…

    113 GitHub stars~1.4k tokensUpdated today
    Auto-check passed
  • Pure Lsp Connect

    finos/legend-engine

    Attaches the pure-lsp HTTP bridge to an LSP daemon that is ALREADY running (e.g.

    113 GitHub stars~1.3k tokensUpdated today
    Auto-check passed
  • Pure Lsp Execute

    finos/legend-engine

    Runs exactly ONE existing Pure function by path (signature, mangled id or bare path) through the running Legend Pure LSP bridge, without a go() wrapper, and returns its typed value in returnValue.

    113 GitHub stars~883 tokensUpdated today
    Auto-check passed
  • Pure Lsp Execute Parallel

    finos/legend-engine

    Runs 2 to 30 Pure functions or tests concurrently on the warm LSP daemon via pure-lsp execute-parallel, or every test in a package or .pure file with --package/--source.

    113 GitHub stars~927 tokensUpdated today
    Auto-check passed

Questions about Pure Backend Start

What does Pure Backend Start do?

Starts (or confirms) the standalone legend-engine backend - engine Server + H2 + local metadata server on fixed ports 9095/9092 - so a Pure LSP started with -Dlegend.test. Pure Backend Start is an agent skill from finos/legend-engine.test.

When should I use Pure Backend Start?

Pure Backend Start fits situations like: the user wants to run Pure queries through the real engine; through plan-gen; via the backend server; reproduce compiled-mode behaviour the interpreted engine cannot.

How do I install Pure Backend Start in Claude Code?

Run `npx skills add finos/legend-engine --skill pure-backend-start -a claude-code`. Or copy the skill folder (plugins/pure-dev/skills/pure-backend-start in finos/legend-engine) into .claude/skills/pure-backend-start in your project. Claude Code loads it when a task matches its description.

How do I install Pure Backend Start in Codex?

Run `npx skills add finos/legend-engine --skill pure-backend-start -a codex`. Or copy the skill folder (plugins/pure-dev/skills/pure-backend-start in finos/legend-engine) into .agents/skills/pure-backend-start in your project. Codex loads it when a task matches its description.

Can I use Pure Backend Start 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 finos/legend-engine --skill pure-backend-start -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/pure-backend-start, .gemini/skills/pure-backend-start, .github/skills/pure-backend-start and .opencode/skills/pure-backend-start in your project.

What does Pure Backend Start need to run?

Going by SKILL.md and its folder, Pure Backend Start needs the command-line tools its instructions call (mvn).

Does Pure Backend Start access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Pure Backend Start 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 Pure Backend Start use?

Pure Backend Start 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 Pure Backend Start use?

About 1.2k tokens (SKILL.md is roughly 5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Pure Backend Start?

Skills that share tags, products or a category with Pure Backend Start: Start (Donchitos/Claude-Code-Game-Studios, 26k stars), Developing Standalone (TriliumNext/Trilium, 38k stars), Session Start Hook (asgeirtj/system_prompts_leaks, 69k stars) and Save As Standalone HTML (asgeirtj/system_prompts_leaks, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Pure Backend Start?

finos (a GitHub organization) maintains it in finos/legend-engine, which has 113 GitHub stars. The repository holds 15 skills in this directory. The repository was last updated on October 8, 2026.

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