K-Dense-AI/scientific-agent-skills / skills/exa-search
exa-search
VerifiedSearches scientific and technical web content with Exa and extracts page or PDF text from URLs in batches. Supports scholarly discovery with the publication category and academic domain filters. Applies to requests to search the web, look up current research, fetch a page, or extract an article using Exa.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.2 KB |
| references/web-extract.md | markdown | 3.4 KB |
| references/web-search.md | markdown | 6.5 KB |
| scripts/exa_extract.py | script | 4.4 KB |
| scripts/exa_search.py | script | 6.4 KB |
references/web-search.md
View on GitHub# Web Search
Search the web for: $ARGUMENTS
## Command
Choose a short, descriptive filename based on the query (e.g., `ai-chip-news`, `crispr-off-target`). Use lowercase with hyphens, no spaces.
```bash
uv run "$SKILL_PATH/scripts/exa_search.py" "$ARGUMENTS" \
--text --highlights \
-o "$FILENAME.json"
```
`$SKILL_PATH` is the path to this skill directory. The `-o` flag saves the full results to a JSON file so follow-up questions can reuse them without re-querying.
**Search type selection** — `--type` controls retrieval mode:
| Mode | When to use |
|---|---|
| `auto` (default) | Exa's general-purpose search. Use this unless you have a reason not to. |
| `fast` | Low-latency interactive search. |
| `instant` | Minimum response time, trading some depth for speed. |
| `deep-lite` | Lightweight research with lower latency than deep. |
| `deep` | Multi-step research for complex queries. |
| `deep-reasoning` | Higher-effort reasoning for complex analysis. |
**Content modes** — add any combination:
- `--text` returns full-text content per result
- `--highlights` returns the most relevant passages (good signal-to-noise, lower token cost than full text)
Default to `--highlights` for broad searches (less text to process). Add `--text` only when you need to quote or extract in detail.
**Filtering options** — Exa supports rich filtering via the SDK:
- `--start-published-date 2026-01-01T00:00:00Z` / `--end-published-date 2026-10-01T00:00:00Z` for ISO 8601 publication-date bounds
- `--include-domains domain1.com,domain2.com` to restrict to an allowlist
- `--exclude-domains spam.com,low-quality.com` to drop a blocklist
- `--category publication` to bias toward scholarly content (also: `company`, `news`, `personal site`, `financial report`, `people`; other strings are category hints)
- `--user-location US` for locale-specific results
`--num-results` accepts 1–100 (individual modes may impose lower limits). There
is no documented offset/cursor pagination; reformulate a query or partition by
supported filters when more coverage is needed. Neither method establishes an
exhaustive literature search. `company` and `people` reject publication-date and
exclude-domain filters; the wrapper catches these combinations before calling.
Use `--include-domains github.com` instead of the deprecated `github` category.
Domain filters also support paths and wildcard subdomains. No content flag
means metadata only, despite the SDK's default text retrieval.
## Academic source strategy
For scientific or technical queries, Exa has two strong levers:
### 1. Use `--category publication`
```bash
uv run "$SKILL_PATH/scripts/exa_search.py" "$ARGUMENTS" \
--category publication \
--text --highlights \
-o "$FILENAME-academic.json"
```
This selects the current scholarly-publication category (papers, preprints, and journal articles). Verify peer review, publication metadata, and relevance on the source; the category does not establish evidence quality.
### 2. Restrict to scholarly domains
For stricter academic filtering, combine the category with an explicit domain allowlist:
```bash
uv run "$SKILL_PATH/scripts/exa_search.py" "$ARGUMENTS" \
--category publication \
--include-domains "arxiv.org,biorxiv.org,medrxiv.org,pubmed.ncbi.nlm.nih.gov,nature.com,science.org" \
--text --highlights \
-o "$FILENAME-academic.json"
```
### Two-pass pattern for comprehensive coverage
Run **both** an academic-focused search and an unrestricted one, then merge with academic sources first:
1. Academic pass: `--category publication` with the scholarly domain allowlist above.
2. General pass: the standard command without `--category` or `--include-domains`, to catch relevant non-academic sources (news coverage, lab blogs, institutional pages).
Merge results, leading with academic sources. If the query is clearly non-scientific, skip the academic pass.
**When to use the two-search pattern:** Any query involving scientific claims, medical information, research findings, technical mechanisms, statistical data, or anything where primary literature would be more reliable than secondary reporting.
## Parsing results
Parse the JSON output. Each result includes:
- `title`, `url`, `published_date`, `author`
- `text` (if `--text`) and `highlights` (if `--highlights`)
- Legacy compatibility fields `score`, `highlight_scores`, and top-level
`autoprompt_string` may be null/empty; never require them or infer evidence
quality from a retrieval score. Exa's changelog says `highlightScores` was
removed in May 2026 although the generated reference still lists it.
Missing author/date/text metadata is normal. The wrapper exports a selected
subset of the SDK response, not raw API JSON or structured publication entities.
**Snippet fallback** — any combination of content fields may be present. Cascade through them: prefer `highlights` (tight, pre-selected passages), fall back to a truncated slice of `text`. Never assume exactly one is present.
## Response format
**CRITICAL: Every claim must have an inline citation.** Use markdown links pulling only from the JSON output. Never invent or guess URLs.
For academic sources, use author-year citation style where metadata is available:
- Academic: [Smith et al., 2025](url) or [Smith & Jones, 2024](url)
- Non-academic: [Source Title](url)
Synthesize a response that:
- Leads with findings from peer-reviewed or preprint sources when available
- Clearly distinguishes between claims backed by primary research vs. secondary reporting
- Includes specific facts, names, numbers, dates
- Cites every fact inline — do not leave any claim uncited
- Organizes by theme if multiple topics
- Notes the evidence quality (e.g., "a randomized controlled trial found..." vs. "a blog post reports...")
**End with a Sources section** listing every URL referenced, grouped by type:
```
Sources:
Academic / Peer-reviewed:
- [Smith et al., 2025 — Title of Paper](https://doi.org/...) (Nature, 2025)
- [Jones & Lee, 2024 — Title of Paper](https://arxiv.org/...) (arXiv preprint)
Other:
- [Source Title](https://example.com/article) (Feb 2026)
```
This Sources section is mandatory. Do not omit it. If no academic sources were found, report the search coverage and filters used; do not infer that the topic has not been studied.
After the Sources section, mention the output file path (`$FILENAME.json`) so the user knows it's available for follow-up questions.
Verified against the [Search API](https://exa.ai/docs/reference/search) and [changelog](https://exa.ai/docs/changelog) on 2026-09-30; authenticated commands are illustrative.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| canvas-designanthropics/skillsCreate beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| docxanthropics/skillsUse this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pdfanthropics/skillsUse this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pptxanthropics/skillsUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates (.potx), layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx or .potx filename, regardless of what they plan to do with the content afterward. If a .pptx or .potx file needs to be opened, created, or touched, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |