CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

exa-search

Searches 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.

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

67%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, mostly actionable overview: routing is unambiguous, setup is executable end-to-end, and detail is correctly pushed to one-level-deep reference files that all exist. The main costs are review-provenance and time-stamped detail inline in SKILL.md that adds tokens without aiding execution, plus some API-scope detail that belongs in the references.

Suggestions

Move the review-provenance statements ("Reviewed the... on 2026-09-30", "no paid live calls were made") and SDK-scope detail out of SKILL.md or into a short 'Verified against' note in the reference files, keeping the overview operational.

Relocate the "Extraction limits and freshness" specifics (URL batch limits, --max-age-hours semantics) to references/web-extract.md, which already covers them, leaving only a one-line pointer in the body.

Consolidate version/date-sensitive details (SDK versions, review dates) into a single compatibility/deprecation note so time-sensitive facts don't penalize the body's conciseness.

DimensionReasoningScore

Conciseness

The routing table and setup/auth commands are tight and operational, but sections like "Reviewed the [Search API]... on 2026-09-30" and "no paid live calls were made during this review" are review-provenance notes rather than instructions, and time-sensitive version/date details sit inline rather than in a deprecation/old-patterns section. This fits anchor 3 (mostly efficient but includes unnecessary explanation that could be trimmed) more than anchor 4's "minor instances".

3 / 5

Actionability

Copy-paste-ready commands are given for setup ("uv run \"$SKILL_PATH/scripts/exa_search.py\" --help", "uv pip install \"exa-py>=2.23.0,<3\"", the dotenv invocation, "export EXA_API_KEY=\"your-key\"") and error-reporting rules are concrete. Common-case search/extract invocations are delegated to the references rather than shown inline, leaving minor gaps versus anchor 5.

4 / 5

Workflow Clarity

The flow is clearly sequenced: route via the decision table, "Read the corresponding reference file for detailed instructions before running commands", then an ordered setup (check .env → dotenv or export → verify) with an explicit checkpoint ("Use --help to verify installation and CLI parsing; it does not require an API key") and batch-error guidance ("report failures without inventing a cause"). Anchor 5's full validate→fix→retry feedback loops are absent, so anchor 4 fits best.

4 / 5

Progressive Disclosure

Structure is good: a routing table, decision guide, and "Files in this skill" listing navigate cleanly to references/web-search.md and references/web-extract.md, both of which exist, are clearly signaled, and are one level deep with no nested .md references. However, "Extraction limits and freshness" and "Verified API and SDK scope" inline reference-level detail that belongs in the references, matching anchor 4 (minor organization gaps) rather than anchor 5's fully appropriate split.

4 / 5

Total

15

/

20

Passed

Description

88%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: third-person, concrete about both capabilities, with an explicit trigger clause covering natural request phrasings. Its only weaknesses are a handful of missing trigger synonyms and mild overlap risk with general-purpose web search/fetch skills.

Suggestions

Add one or two more natural trigger variations (e.g., "web search", "research a topic", "read this link/URL") to the when-clause for broader trigger coverage.

Sharpen the distinctiveness against built-in web tools, e.g., by signaling when to prefer Exa (scholarly/semantic or batch PDF extraction) over a generic web search.

DimensionReasoningScore

Specificity

"Searches scientific and technical web content with Exa", "extracts page or PDF text from URLs in batches", and "publication category and academic domain filters" list multiple concrete actions covering both core capabilities plus scholarly modes. Coverage is comprehensive for the skill's scope, matching the anchor-5 example rather than anchor 4, which expects minor gaps.

5 / 5

Completeness

The what is explicit (search + batch page/PDF extraction with scholarly filters) and the when is explicit via "Applies to requests to search the web, look up current research, fetch a page, or extract an article using Exa" — concrete trigger phrases, matching anchor 5. Not 4, because the when-clause is explicit and specific rather than improvable.

5 / 5

Trigger Term Quality

Natural trigger phrases like "search the web", "look up current research", "fetch a page", and "extract an article" appear alongside "PDF" and "URLs". A few natural variations users would say (e.g., "web search", "research a topic", "read this link") are missing, so it fits anchor 4 (good coverage, a few terms missing) rather than anchor 5's comprehensive synonym coverage.

4 / 5

Distinctiveness Conflict Risk

The "using Exa" qualifier and the scholarly/publication niche give it a clear identity distinct from sibling skills. However "search the web" and "fetch a page" are broad triggers that overlap generic web-search/fetch skills, matching anchor 4 (mostly distinct, minor overlap with closely related skills) rather than anchor 5's minimal conflict risk.

4 / 5

Total

18

/

20

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_field

'metadata' should map string keys to string values

Warning

relative_links

Relative link issues: 1 missing

Warning

Total

14

/

16

Passed

Repository

K-Dense-AI/scientific-agent-skills

Path

skills/exa-search/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.