Agent skill

Baleen Multihop

by VectorSpaceLab in VectorSpaceLab/AREX-Skill

Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats.

MITAuto-check passed

Install Baleen Multihop

skills CLI
$ npx skills add VectorSpaceLab/AREX-Skill --skill baleen-multihop -a claude-code

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

GitHub CLI
$ gh skill install VectorSpaceLab/AREX-Skill baleen-multihop --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/VectorSpaceLab/AREX-Skill.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/repositories/repo-skills/colbert/sub-skills/baleen-multihop .claude/skills/baleen-multihop && 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
baleen-multihop
GitHub stars
330
Token cost
~921 tokens
SKILL.md length
376 words
Files
5 (incl. scripts, references)
Skills in repo
159
Repo updated
First seen
Licence
MIT

At a glance

Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats.

  • Tasks mention Baleen
  • SKILL.md covers Route here when, Route elsewhere when, Safe operating stance and Common actions, plus 1 more section
  • Runs Python scripts from its folder; calls python
  • Multi-hop retrieval

What it does

Baleen Multihop is an agent skill from VectorSpaceLab/AREX-Skill. Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats. Use when tasks mention Baleen, multi-hop retrieval, condensed retrieval, HopSearcher, Condenser, collectionX, or L1/L2 condenser checkpoints.

Its SKILL.md is about 920 tokens, which your agent loads only when the skill is triggered. The skill folder holds 6 other files, including scripts and reference files (for example `references/api-reference.md`, `references/baleen-workflows.md` and `references/troubleshooting.md`).

The repository describes itself as: A Skill Library for Automated Machine Learning. The licence is MIT.

When your agent uses it

  • Tasks mention Baleen
  • Multi-hop retrieval
  • Condensed retrieval
  • L1/L2 condenser checkpoints

Example prompts

  • “/baleen-multihop”

Requirements

  • Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit ac3fe1a. 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

    Ships 1 file in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python

    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

Baleen Multihop loads about 921 tokens when it runs, and up to ~5.5k if it reads all its reference files. Until then it costs about 82 tokens; SKILL.md has 376 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~82
When it runs · the whole SKILL.md, loaded when a task matches
~921
With references · SKILL.md plus every file in references/, read only if the agent opens them
~5.5k

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); the scripts in this folder are not scanned.

SKILL.md

The full file from VectorSpaceLab/AREX-Skill at commit ac3fe1a, republished under its MIT licence (© VectorSpaceLab). 376 words, ~921 tokens.

Download SKILL.mdSave it as .claude/skills/baleen-multihop/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.
name
baleen-multihop
description
Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats. Use when tasks mention Baleen, multi-hop retrieval, condensed retrieval, HopSearcher, Condenser, collectionX, or L1/L2 condenser checkpoints.
disable-model-invocation
true
metadata.disco-role
operating
license
MIT

Baleen Multi-hop

Use this sub-skill for ColBERT's optional Baleen multi-hop retrieval components, not for ordinary one-hop ColBERT indexing or search. Baleen composes a ColBERT Searcher variant with an Electra-based Condenser to retrieve passages, select sentence-level facts, and feed those facts as context into later hops.

Route here when

  • The user asks about Baleen, multi-hop retrieval, condensed retrieval, hop-aware search, HopSearcher, Condenser, collectionX, or L1/L2 condenser checkpoints.
  • The code path involves baleen.engine.Baleen, baleen.hop_searcher.HopSearcher, baleen.condenser.condense.Condenser, baleen.condenser.model.ElectraReader, or baleen.condenser.tokenization.AnswerAwareTokenizer.
  • The user has, or needs to reason about, a built ColBERT index plus Baleen-specific sentence JSONL and condenser checkpoint artifacts.
  • The safe answer is to inspect imports/signatures, explain prerequisites, plan an adaptation, or diagnose why a full Baleen run cannot be validated locally.

Route elsewhere when

  • Use indexing-and-search for normal Indexer, Searcher, collection TSVs, one-hop search, or ranking files.
  • Use training-and-distillation for training or fine-tuning ColBERT retriever checkpoints.
  • Use data/evaluation guidance for non-Baleen data conversion and metrics; keep only collectionX sentence-format notes here.
  • Use modeling/tokenization guidance for ColBERT checkpoint internals unless the tokenizer issue is specifically Baleen's answer-aware condenser tokenizer.

Safe operating stance

Baleen is bundled as a research-style extension. Imports and source signatures can be inspected on CPU, but end-to-end multi-hop retrieval usually needs local ColBERT index artifacts, a retrieval checkpoint, two Baleen Electra reader checkpoints, a compatible sentence-level collectionX file, and often CUDA-capable devices. Do not promise full execution from package import success alone.

Start by confirming artifacts and device constraints. If the user is CPU-only or lacks checkpoints/data, run the import inspector, explain what is available, and provide a static adaptation plan rather than attempting retrieval.

Show full SKILL.md (111 more words)Show less

Common actions

  • Use references/api-reference.md when you need source-backed class/function behavior, signatures, and composition notes.
  • Use references/baleen-workflows.md when planning a safe multi-hop flow, validating artifact shape, or explaining how facts become context.
  • Use references/troubleshooting.md for install/import, optional dependency, CUDA/device, checkpoint, collectionX, API misuse, and depth/hop failures.
  • Use scripts/inspect_baleen_imports.py when you need deterministic import/signature diagnostics without loading checkpoints, building indexes, downloading datasets, or running retrieval.

Minimal diagnostic

From an environment where colbert-ai is installed and colbert is importable:

bash
python scripts/inspect_baleen_imports.py --tiny-fixture

Add --json for machine-readable output. Treat a passing report as evidence that Baleen import surfaces and loader behavior are available, not as evidence that the user's checkpoints, GPU, index, or full multi-hop workflow will run.

© VectorSpaceLab, MIT. 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 4 other files (scripts, references) in skills/repositories/repo-skills/colbert/sub-skills/baleen-multihop of VectorSpaceLab/AREX-Skill.

  • SKILL.md
  • references/api-reference.md
  • references/baleen-workflows.md
  • references/troubleshooting.md
  • scripts/inspect_baleen_imports.py

Open the folder on GitHubat commit ac3fe1a

Compare with similar skills

Baleen Multihop 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.

Baleen Multihop compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Baleen Multihop this skillVectorSpaceLab/AREX-Skill330—~921Automated safety check: PassMIT
Optionsasgeirtj/system_prompts_leaks69k—~918Automated safety check: PassCC0-1.0
Browser Extension Buildersickn33/agentic-awesome-skills47k2 repos~2.3kAutomated safety check: PassMIT
Browser Extension Launchsickn33/agentic-awesome-skills47k1 repos~1.4kAutomated safety check: PassMIT
Browser Extension Reversesickn33/agentic-awesome-skills47k1 repos~695Automated safety check: PassMIT
Iterative Retrievalaffaan-m/ECC276k7 repos~1.6kAutomated safety check: PassMIT

Similar skills

  • Options

    asgeirtj/system_prompts_leaks

    Present multiple design options as a vertical stack of anchored turns

    69k GitHub stars~918 tokensUpdated yesterday
    Auto-check passed
  • Browser Extension Builder

    sickn33/agentic-awesome-skills

    Expert in building browser extensions that solve real problems - Chrome, Firefox, and cross-browser extensions.

    47k GitHub starsUsed in 2 repos~2.3k tokens
    DevelopmentAuto-check passed
  • Browser Extension Launch

    sickn33/agentic-awesome-skills

    Builds, tests, packages, and prepares Chrome extensions for store launch from a plain-language idea; use for new extensions, fixes, releases, and submission recovery.

    47k GitHub starsUsed in 1 repo~1.4k tokens
    DevelopmentAuto-check passed
  • Browser Extension Reverse

    sickn33/agentic-awesome-skills

    Authorized reverse engineering of Chrome/Firefox extensions: manifest analysis, background workers, content scripts, and extension-based credential or data-exposure research.

    47k GitHub starsUsed in 1 repo~695 tokens
    SecurityAuto-check passed
  • Pattern for progressively refining context retrieval to solve the subagent context problem.

    276k GitHub starsUsed in 7 repos~1.6k tokens
    Agent WorkflowsAuto-check passed
  • Browser Extension Reverse

    zhaoxuya520/reverse-skill

    A skill your agent uses for authorized reverse engineering of browser extensions (Chrome/Firefox) including manifest analysis, background workers, and extension-based credential or traffic logic…

    40k GitHub starsUsed in 2 repos~366 tokens
    SecurityAuto-check passed

More from VectorSpaceLab/AREX-Skill

All 159 skills in this repo
  • Agent Lightning

    VectorSpaceLab/AREX-Skill

    Use this repo skill for Agent Lightning package tasks: authoring trainable agents, tracing rewards and spans, running LightningStore/Trainer loops, using agl CLI services, choosing examples, and…

    330 GitHub stars~1.4k tokensUpdated 1 mo ago
    Auto-check passed
  • Agent Tools

    VectorSpaceLab/AREX-Skill

    A skill your agent uses when configuring LiteLLM for MCP tools, A2A agents, Claude Code/Cursor agent gateway traffic, MCP auth/OAuth, tool permissions, semantic filtering, or agent-specific proxy…

    330 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed
  • Agents And Awel

    VectorSpaceLab/AREX-Skill

    Build and debug DB-GPT agents, tools, skills, teams, and AWEL workflows, including deterministic local DAG runs and HTTP-trigger topology without assuming an LLM, credential, or external service.

    330 GitHub stars~2.3k tokensUpdated 1 mo ago
    Auto-check passed
  • Agents And Middleware

    VectorSpaceLab/AREX-Skill

    Work on the actively maintained LangChain v1 agent package: initchatmodel, createagent, structured output, tools, middleware, embeddings initialization, provider routing, and agent runtime…

    330 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed
  • Agents Workflows

    VectorSpaceLab/AREX-Skill

    A skill your agent uses for giskard.agents async chat workflows, tools, prompt templates, structured outputs, retries, rate limiting, embeddings, and optional LiteLLM backend.

    330 GitHub stars~500 tokensUpdated 1 mo ago
    Auto-check passed
  • Alphafold3

    VectorSpaceLab/AREX-Skill

    A skill your agent uses for AlphaFold 3 input preparation, prediction command planning, output interpretation, and Python API inspection.

    330 GitHub stars~1.2k tokensUpdated 1 mo ago
    Auto-check passed

Questions about Baleen Multihop

What does Baleen Multihop do?

Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats. Baleen Multihop is an agent skill from VectorSpaceLab/AREX-Skill. Optional ColBERT Baleen multi-hop retrieval extension: HopSearcher context search, Baleen engine loops, Condenser fact selection, collectionX format, and artifact caveats.

When should I use Baleen Multihop?

Baleen Multihop fits situations like: tasks mention Baleen; multi-hop retrieval; condensed retrieval; L1/L2 condenser checkpoints.

How do I install Baleen Multihop in Claude Code?

Run `npx skills add VectorSpaceLab/AREX-Skill --skill baleen-multihop -a claude-code`. Or copy the skill folder (skills/repositories/repo-skills/colbert/sub-skills/baleen-multihop in VectorSpaceLab/AREX-Skill) into .claude/skills/baleen-multihop in your project. Claude Code loads it when a task matches its description.

How do I install Baleen Multihop in Codex?

Run `npx skills add VectorSpaceLab/AREX-Skill --skill baleen-multihop -a codex`. Or copy the skill folder (skills/repositories/repo-skills/colbert/sub-skills/baleen-multihop in VectorSpaceLab/AREX-Skill) into .agents/skills/baleen-multihop in your project. Codex loads it when a task matches its description.

Can I use Baleen Multihop 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 VectorSpaceLab/AREX-Skill --skill baleen-multihop -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/baleen-multihop, .gemini/skills/baleen-multihop, .github/skills/baleen-multihop and .opencode/skills/baleen-multihop in your project.

What does Baleen Multihop need to run?

Going by SKILL.md and its folder, Baleen Multihop needs Python for the scripts in its folder and the command-line tools its instructions call (python). Our summary lists: Python 3.

Does Baleen Multihop 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 Baleen Multihop 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. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Baleen Multihop use?

Baleen Multihop is published under the MIT licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Baleen Multihop use?

About 921 tokens (SKILL.md is roughly 3.7k 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 4.6k tokens, read only when the agent opens those files.

What are the alternatives to Baleen Multihop?

Skills that share tags, products or a category with Baleen Multihop: Options (asgeirtj/system_prompts_leaks, 69k stars), Browser Extension Builder (sickn33/agentic-awesome-skills, 47k stars), Browser Extension Launch (sickn33/agentic-awesome-skills, 47k stars) and Browser Extension Reverse (sickn33/agentic-awesome-skills, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Baleen Multihop?

VectorSpaceLab (a GitHub organization) maintains it in VectorSpaceLab/AREX-Skill, which has 330 GitHub stars. The repository holds 159 skills in this directory. The repository was last updated on September 3, 2026.

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