Asta Skill

Tool Map — Intent → Asta Tool

All tools accept date-range filters and field selection — pass them whenever the user's intent constrains scope (e.g., "recent", "since 2022", "at NeurIPS").

⚠️ fields parameter — avoid context blowups

get_paper / get_paper_batch accept a fields string. Never request citations or references via fields — a single highly-cited paper (e.g. Attention Is All You Need) returns 200k+ characters and will overflow the agent's context window. Use the dedicated get_citations tool instead (it paginates).

Safe default fields for get_paper:

title,year,authors,venue,tldr,url,abstract

Add journal, publicationDate, fieldsOfStudy, isOpenAccess only when needed.

Workflow Patterns

Pattern 1 — Topic Discovery

1. search_papers_by_relevance(query, year="2022-", venue=?) → initial hits
2. Rank/present top N by citationCount + recency
3. Offer follow-ups: get_citations on the most influential, or snippet_search for specific claims

Pattern 2 — Seed-Paper Expansion

1. get_paper(DOI|arXiv|...) → verify seed
2. get_citations(paperId) → forward expansion
3. Optionally search_papers_by_relevance with seed title terms for sideways discovery
4. Deduplicate by paperId before presenting

Pattern 3 — Author Deep-Dive

1. search_authors_by_name(name) → pick correct profile (disambiguate by affiliation)
2. get_author_papers(authorId) → full publication list
3. Filter client-side by topic keywords or date

Pattern 4 — Evidence Retrieval

1. snippet_search(claim_query) → find passages making/supporting a claim
2. For each hit, optionally get_paper(id) for full metadata

Output & Interaction Rules

• Always report total count and which tool was used.
• Present top 10 as a table (title, year, venue, citations), then details for the most relevant.
• If the user writes in Chinese, present summaries in Chinese; keep titles in original language.
• After results, offer: Details / Refine / Citations / Snippet / Export / Done.

Critical Rules

• Prefer batched intent over ping-pong. If the user's question needs two independent lookups, issue them as parallel MCP tool calls in one turn, not sequentially.
• Never guess IDs. If a user gives a fuzzy title, use search_paper_by_title before get_paper.
• Respect rate limits. An API key buys higher limits but not unlimited — stop expanding citation graphs beyond what the user asked for.
• Do not fabricate fields. If Asta returns null abstract or venue, say so rather than inventing.

Relationship to semanticscholar-skill

Both wrap the Semantic Scholar corpus, but target different runtimes:

Use asta-skill when the host agent supports MCP; fall back to semanticscholar-skill for scripted/pipeline work.

Installation

Set ASTA_API_KEY in your shell first:

export ASTA_API_KEY="..."   # request at https://share.hsforms.com/1L4hUh20oT3mu8iXJQMV77w3ioxm

Claude Code

claude mcp add asta \
  --transport http \
  --url https://asta-tools.allen.ai/mcp/v1 \
  --header "x-api-key: $ASTA_API_KEY"

Or edit ~/.claude.json / .mcp.json:

{
  "mcpServers": {
    "asta": {
      "type": "http",
      "url": "https://asta-tools.allen.ai/mcp/v1",
      "headers": { "x-api-key": "${ASTA_API_KEY}" }
    }
  }
}

Codex CLI

Edit ~/.codex/config.toml:

[mcp_servers.asta]
type = "http"
url = "https://asta-tools.allen.ai/mcp/v1"
headers = { "x-api-key" = "${ASTA_API_KEY}" }

Windsurf / Cursor / Hermes / other MCP clients

Add to the client's MCP server config file:

{
  "mcpServers": {
    "asta": {
      "serverUrl": "https://asta-tools.allen.ai/mcp/v1",
      "headers": { "x-api-key": "<YOUR_API_KEY>" }
    }
  }
}

LM Studio

LM Studio 0.3.17+ supports remote MCP servers. Edit ~/.lmstudio/mcp.json (macOS/Linux) or %USERPROFILE%\.lmstudio\mcp.json (Windows) — or in the app: Program tab → Install > Edit mcp.json:

{
  "mcpServers": {
    "asta": {
      "url": "https://asta-tools.allen.ai/mcp/v1",
      "headers": { "x-api-key": "<YOUR_API_KEY>" }
    }
  }
}

Only models with "Tool Use: Supported" in LM Studio's model loader will be able to call Asta tools. Recommended: Qwen 2.5 / 3 Instruct (7B+), Llama 3.1 / 3.3 Instruct (8B+), Mistral / Mixtral Instruct.

OpenClaw

Install this skill into ~/.openclaw/skills/asta-skill/ and register the MCP server in your OpenClaw config using the same URL + x-api-key header pattern. The skill's frontmatter declares ASTA_API_KEY as required via metadata.openclaw.requires.env.

Verification

After installation, ask the agent: "Use Asta to look up the paper with DOI 10.48550/arXiv.1706.03762." A successful call returns the "Attention Is All You Need" paper metadata. If the agent reports no Asta tools, the MCP server is not registered — re-check the config file path and restart the host.

Troubleshooting