Bmad Deep Recon
delorenj/mcp-server-trello
Decision-grade research, three ways: draft a deep-research prompt for the user to run in their own tool (ChatGPT, Gemini, Grok, Perplexity, …), process a finished research report — file it, distill…
Runs evidence-aware, multi-provider research with the Librarium v2 CLI.
Install Librarium
$ npx skills add jkudish/librarium --skill librarium -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install jkudish/librarium librarium --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .claude/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jkudish/librarium --skill librarium -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install jkudish/librarium librarium --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .agents/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jkudish/librarium --skill librarium -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install jkudish/librarium librarium --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .cursor/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jkudish/librarium --skill librarium -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install jkudish/librarium librarium --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .gemini/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install jkudish/librarium librariumInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add jkudish/librarium --skill librarium -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .github/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add jkudish/librarium --skill librarium -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install jkudish/librarium librarium --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "librarium" agent skill from https://github.com/jkudish/librarium/tree/main into .opencode/skills/librarium/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "librarium", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
librarium- GitHub stars
- 134
- Token cost
- ~1.9k tokens
- SKILL.md length
- 891 words
- Files
- 622 (incl. scripts)
- Skills in repo
- 1
- Repo updated
- First seen
- Licence
- MIT
At a glance
Runs evidence-aware, multi-provider research with the Librarium v2 CLI.
- Works in 7 steps: Query Analysis → Provider Selection → Dispatch → …
- Competitive research
- SKILL.md covers Prerequisites, 7-Phase Research Workflow, Key Commands and MCP Server, plus 2 more sections
- Runs Shell scripts from its folder; calls npm and claude
What it does
Librarium is an agent skill from jkudish/librarium. Runs evidence-aware, multi-provider research with the Librarium v2 CLI. Use for deep research, competitive research, answer-engine visibility checks, or questions needing grounded multi-source coverage.
Its SKILL.md is about 1.9k tokens, which your agent loads only when the skill is triggered. The skill folder holds 626 other files, including scripts (for example `.github/FUNDING.yml`, `.github/ISSUE_TEMPLATE/bug_report.yml` and `.github/ISSUE_TEMPLATE/config.yml`). Compatibility notes: Requires Node.js 22.12 or newer and the Librarium 2.x CLI.
It sits in Research & Science, covering Deep research. It works with Google Gemini, OpenAI and Perplexity. The repository describes itself as: Multi-provider deep research CLI — fans out queries to multiple search/AI APIs in parallel. The licence is MIT.
When your agent uses it
- Competitive research
- Answer-engine visibility checks
- Questions needing grounded multi-source coverage
Example prompts
- “/librarium”
Requirements
- Node.js
- A Bash shell
- Compatibility (from SKILL.md): Requires Node.js 22.12 or newer and the Librarium 2.x CLI.
Workflow steps
7 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit 2137a9a. 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/ (Shell, from the files we listed), which the agent can run.
Shell commands in SKILL.md call:
npmclaude
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.
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.
Compatibility
Requires Node.js 22.12 or newer and the Librarium 2.x CLI.
From compatibility in the SKILL.md frontmatter.
Context cost
Librarium loads about 1.9k tokens when it runs. Until then it costs about 53 tokens; SKILL.md has 891 words of instructions outside code blocks.
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 jkudish/librarium at commit 2137a9a, republished under its MIT licence (© jkudish). 891 words, ~1,947 tokens.
Download SKILL.mdSave it as
.claude/skills/librarium/SKILL.md (or your agent's skills folder). This skill also uses 621 other files; get the full folder from GitHub.- name
- librarium
- description
- Runs evidence-aware, multi-provider research with the Librarium v2 CLI. Use for deep research, competitive research, answer-engine visibility checks, or questions needing grounded multi-source coverage.
- compatibility
- Requires Node.js 22.12 or newer and the Librarium 2.x CLI.
Librarium -- Multi-Provider Deep Research
Run research queries through Librarium’s v2 public provider/profile catalog. There are 34 built-in providers and 41 implemented profiles. Preserve the profile and collection provenance when reporting results; do not turn source counts or agreement into a confidence claim.
Prerequisites
Before the first query in an orb, run librarium --version and require major
version 2. The Librarium repository's .agents/setup installs its built v2
checkout globally. In other orbs, install the published v2 package when it is
available:
npm install -g 'librarium@^2'Do not silently fall back to Librarium 1.x or install a mutable Git branch. If
npm has no v2 release yet, report that distribution blocker instead. Configure
available API keys with librarium init --auto, then inspect the usable matrix
offline with librarium doctor and librarium ls --json. This confirms only
configuration and credential presence, not authentication or connectivity, and
does not load configured custom provider code. Use librarium doctor --live
only when the user explicitly wants live connectivity tests; it loads trusted
custom providers, makes provider network requests, and may incur charges. Never
print secret values.
7-Phase Research Workflow
Phase 1: Query Analysis
Analyze the user's research question. Determine:
- Is this a technical, business, or general knowledge query?
- Which built-in workflow is best suited? (
quickfor a curated low-latency set,deepfor research-report profiles,visibilityfor an explicit nine-perspective comparison,allfor catalog-derived coverage). Usecustom:<name>for a user-authored group. - What execution mode? (
syncfor quick queries,mixedfor deep research)
Phase 2: Provider Selection
Select providers based on query type:
- Technical queries: Start with
quick; add named durable profiles ordeeponly when the question warrants the latency and possible spend. - Quick facts: Use
quickgroup (AI-grounded only, fast) - Competitive research: Use
allonly after confirming its configured, credentialed membership and cost exposure. - Answer visibility: Use
visibilityonly when you deliberately want six SearchAPI-collected consumer-surface observations compared with three first-party API baselines. Treat the six collection-vendor results as correlated evidence, not independent confirmation. - Specific provider: Use
--providers(canonical IDs or display names, e.g.-p "Exa Search,brave-search")
Phase 3: Dispatch
Run the query:
librarium run "your query here" --group <group> [--mode mixed]Phase 4: Monitor Async Tasks
If a profile is background/durable and was submitted in async mode:
librarium status --waitPhase 5: Retrieve Results
Once the provider is observed complete, retrieve the terminal result:
librarium status --retrievePhase 6: Analyze Output
Read the output files:
summary.md-- Overall research summary with statisticssources.json-- Deduplicated citations ranked by frequency- Individual
{provider}.mdfiles for detailed per-provider results run.json-- Machine-readable manifest
Phase 7: Synthesize
Combine findings from multiple providers into a coherent answer. Record source overlap and provenance, but do not convert a higher citation count into a confidence score: shared collectors and repeated sources can make that evidence correlated.
Show full SKILL.md (445 more words)Show less
Key Commands
| Command | Purpose |
|---|---|
librarium run <query> | Run research query |
librarium run <query> --group quick | Fast AI-grounded search |
librarium run <query> --group deep | Deep research (async) |
librarium run <query> --group visibility | Compare six collected consumer surfaces with three first-party API baselines |
librarium run <query> --group all | Catalog-derived selectable coverage |
librarium answer <query> | Fan out (default quick) and synthesize one grounded, cited answer to answer.md |
librarium run <query> --max-cost 0.50 | Stop launching providers once API-reported cost crosses the budget |
librarium run <query> --yes | Skip the deep-research pre-flight confirm (3+ deep providers) |
librarium status | Check async tasks |
librarium status --wait --retrieve | Wait and fetch Node CLI async results |
librarium live-validation --fixture /absolute/path/to/fixture.json | Replay a strict, network-free canonical fixture |
librarium usage [--days N] [--json] | Aggregate API-reported cost and tokens across past runs |
librarium run <query> --html --open | Run, then open an HTML report |
librarium run <query> --jsonl | Run, then write machine-readable results.jsonl |
librarium browse | Browse past runs interactively |
librarium html [run-dir] | Generate report.html for a run |
librarium jsonl [run-dir] | Generate results.jsonl for a run |
librarium refine <goal> | Tier-tuned query variants, no dispatch |
librarium ls | List providers and status |
librarium doctor | Check provider configuration and credential presence offline |
librarium doctor --live | Test connectivity with provider network requests; charges may apply |
librarium config | Show resolved config |
librarium cleanup [--days N] [--dry-run] | Delete run dirs older than N days (default 30) |
librarium clear [--dry-run] [-i] [--yes] | Delete all run dirs (alias for cleanup --all); -i to pick interactively |
MCP Server
Instead of shelling out to the CLI, agents can drive librarium over the Model Context Protocol with librarium mcp (stdio transport). Register it once with claude mcp add librarium -- librarium mcp, then call the tools: research, get_results, check_async, list_providers, list_groups. The research tool runs the same silent file-writing pipeline as librarium run and returns a compact structured result; fetch full provider markdown with get_results.
Provider Tiers
| Tier | Providers | Speed | Depth |
|---|---|---|---|
| background/durable | exa/research, openai-research/research, gemini-deep/research, perplexity-sonar-deep/research, perplexity-deep-research/research, you-research/research, parallel/research, and valyu/research | Minutes to longer | Persisted handles that can be polled and retrieved |
| inline | Search, grounded, collected surface, and chat profiles | Usually seconds | Immediate response; no durable handle |
Visibility and privacy boundary
The six SearchAPI surface profiles observe consumer-facing ChatGPT, Gemini,
Perplexity, Google AI Mode, Bing Copilot, and Google AI Overview output through
one upstream collector. Treat agreement among them as correlated visibility
evidence, not independent corroboration or parity with a particular logged-in
user. zeroRetention is an account capability: if explicitly configured,
Librarium fails closed when the account rejects it. It makes no broader
retention or privacy guarantee. SerpBase may log search queries for billing,
debugging, abuse prevention, and account logs; no zero-retention mode or fixed
public retention period is documented.
Output Structure
./agents/librarium/{timestamp}-{slug}/
prompt.md, run.json, summary.md, sources.json
{provider}.md, {provider}.meta.json
answer.md (when using `librarium answer`)© jkudish, 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 621 other files (scripts) in the repository root of jkudish/librarium.
- SKILL.md
- .agents/resume
- .agents/setup
- .agents/skills
- .gitattributes
- .github/FUNDING.yml
- .github/ISSUE_TEMPLATE/bug_report.yml
- .github/ISSUE_TEMPLATE/config.yml
- .github/ISSUE_TEMPLATE/feature_request.yml
- .github/ISSUE_TEMPLATE/new_provider.yml
- .github/pull_request_template.md
- .github/scripts/parse-changelog.sh
- .github/workflows/ci.yml
- .github/workflows/release-candidate.yml
- .github/workflows/release.yml
- .gitignore
- … and 606 more
Open the folder on GitHubat commit 2137a9a
Compare with similar skills
Librarium 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.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Librarium this skilljkudish/librarium | 134 | — | ~1.9k | Automated safety check: Pass | MIT | |
| Bmad Deep Recondelorenj/mcp-server-trello | 445 | — | ~2.3k | Automated safety check: Pass | MIT | |
| Deep Research MCP Guidepminervini/deep-research-mcp | 114 | — | ~5.8k | Automated safety check: Pass | MIT | |
| Deep Researchjuanandresgs/claude-ctrl | 193 | — | ~3.1k | Automated safety check: Notes | None | |
| Deep Researchsanjay3290/ai-skills | 432 | 9 repos | ~683 | Automated safety check: Notes | Apache-2.0 | |
| GEO-First SEO Audit Toolzubair-trabzada/geo-seo-claude | 11k | — | ~2.8k | Automated safety check: Notes | MIT |
Similar skills
Deep Research MCP Guide
pminervini/deep-research-mcp
Explains how to run, integrate and debug the deep-research-mcp project through its CLI, Python API or MCP server, with OpenAI, Gemini and DR-Tulu backends.
Deep Research
juanandresgs/claude-ctrl
Multi-model deep research with comparative assessment (OpenAI + Perplexity + Gemini).
Deep Research
sanjay3290/ai-skills
Execute autonomous multi-step research using Google Gemini Deep Research Agent.
GEO-First SEO Audit Tool
zubair-trabzada/geo-seo-claude
Audits a website for AI search visibility across ChatGPT, Claude, Perplexity and Google AI Overviews while checking traditional SEO, schema and E-E-A-T content quality.
Engineering Figure Agent
heyu-233/engineering-figure-agent
A skill your agent uses when the user needs engineering or research-paper figures: system architecture diagrams, algorithm workflows, hardware schematics, benchmark charts, ablation plots, figure…
Related
Works with
Categories
Questions about Librarium
What does Librarium do?
Runs evidence-aware, multi-provider research with the Librarium v2 CLI. Librarium is an agent skill from jkudish/librarium. Runs evidence-aware, multi-provider research with the Librarium v2 CLI.
When should I use Librarium?
Librarium fits situations like: competitive research; answer-engine visibility checks; questions needing grounded multi-source coverage.
How do I install Librarium in Claude Code?
Run `npx skills add jkudish/librarium --skill librarium -a claude-code`. Or copy the skill folder (the jkudish/librarium repository) into .claude/skills/librarium in your project. Claude Code loads it when a task matches its description.
How do I install Librarium in Codex?
Run `npx skills add jkudish/librarium --skill librarium -a codex`. Or copy the skill folder (the jkudish/librarium repository) into .agents/skills/librarium in your project. Codex loads it when a task matches its description.
Can I use Librarium 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 jkudish/librarium --skill librarium -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/librarium, .gemini/skills/librarium, .github/skills/librarium and .opencode/skills/librarium in your project.
What does Librarium need to run?
Going by SKILL.md and its folder, Librarium needs a shell for the scripts in its folder and the command-line tools its instructions call (npm and claude). Our summary lists: Node.js; A Bash shell. Compatibility (from SKILL.md): Requires Node.js 22.12 or newer and the Librarium 2.x CLI..
Does Librarium access the network?
SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Is Librarium 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 Librarium use?
Librarium is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Librarium use?
About 1.9k tokens (SKILL.md is roughly 7.8k 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 Librarium?
Skills that share tags, products or a category with Librarium: Bmad Deep Recon (delorenj/mcp-server-trello, 445 stars), Deep Research MCP Guide (pminervini/deep-research-mcp, 114 stars), Deep Research (juanandresgs/claude-ctrl, 193 stars) and Deep Research (sanjay3290/ai-skills, 432 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Librarium?
jkudish (a GitHub user) maintains it in jkudish/librarium, which has 134 GitHub stars. The repository was last updated on October 9, 2026.
Source: jkudish/librarium on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.