perplexity-search
Skill2 starsChanged 4 months agoMITView on GitHub
Grounded web search & research via Perplexity, keyless and pay-per-call over SELAT. Use when asked to \"search the web for <topic>\", \"what's the latest on <topic>\", \"research <topic> with sources\", \"do a deep-research report on <X>\", or \"give me a grounded answer with citations\". Runs Perplexity's cheap web Search by default, and can escalate to a one-shot Agent answer or an async deep-research report when a plain search isn't enough. Paid per call in USDC (on Base) from the user's own self-custody Circle Agent Wallet — no Perplexity API key, no signup. Dry-run first to see live prices; every price the CLI shows already includes SELAT's ~5% routing markup.
- Installs packages
16 other projects use a version of this.
Adopted by
Most recently taken up by
- BrightGold70/skillstheir own version, first seen Nov 8, 2026
- MarieLynneBlock/arcanum-artifextheir own version, first seen Nov 1, 2026
- parcadei/Continuous-Claudetheir own version, first seen Oct 17, 2026
- garri333/Skillstheir own version, first seen Oct 11, 2026
- wu-yc/LabClawtheir own version, first seen Oct 8, 2026
- parcadei/Continuous-Claude-v2their own version, first seen Oct 8, 2026
Used with
Projects with it also use ena-database, gwas-database and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
garri333Skillena-database
Access European Nucleotide Archive via API/FTP. Retrieve DNA/RNA sequences, raw reads (FASTQ), genome assemblies by accession, for genomics and bioinformatics pipelines. Supports multiple formats.
Also in 7 projects that use this
garri333Skillgwas-database
Query NHGRI-EBI GWAS Catalog for SNP-trait associations. Search variants by rs ID, disease/trait, gene, retrieve p-values and summary statistics, for genetic epidemiology and polygenic risk scores.
Also in 7 projects that use this
- +10 more from garri333
Other versions16 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 16
- parcadei/Continuous-Claude★ 3.9ksame name
- parcadei/Continuous-Claude-v2★ 3.9ksame name
- parcadei/Continuous-Claude-v3★ 3.9ksame name
- wu-yc/LabClaw★ 1.1ksame name
- LeonChaoX/qinyan-academic-skills★ 910same name
- escapeWu/perplexity-ai★ 168same name
- qinyan-ai/qinyan-academic-skills★ 80same name
- ZanderRuss/obsidian-claude★ 19same name
- MarieLynneBlock/arcanum-artifex★ 4same name
- SELAT-AI/selat-skills★ 2same name
6 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository4 other files it gives its agents.
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 39
Every call is real USDC from the user's own Circle Agent Wallet (MPC self-custody — SELAT never holds keys or funds).
- Distinctive: line 53
Never ask for, paste, or handle a private key.
- Very specific: line 115
When one search pass isn't enough, the installed skill documents two higher-tier moves — Agent answer (a one-shot synthesized answer) and …
- Very specific: line 123
Follow the exact request shapes and steps in the installed skill's own docs — its SKILL.md ("Escalations") and references/endpoints.md — …
“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
perplexity-search
A keyless Perplexity research toolkit over SELAT. The default action is a cheap web Search — ranked results with page content and source URLs that you (the agent) synthesize into a cited answer. When a search-and-synthesize pass isn't enough, the skill also documents two escalations — a one-shot Agent answer and an async deep-research report — that you run per the installed skill's own guidance, always confirming the cost first.
It wraps a SELAT skill: a declarative, vetted recipe of paid API calls (no API
keys, no signups) settled in USDC — routed x402 on Base through the SELAT
Router. The selat CLI resolves the vetted endpoints and prints a receipt.
Cost — read this first
- Every call is real USDC from the user's own Circle Agent Wallet (MPC self-custody — SELAT never holds keys or funds).
- Prices and spend limits live in the underlying SELAT skill, not here — the
live 402 quote from
selat skill verify/runis the price source of truth, so this wrapper doesn't restate dollar figures (they'd only drift). - The quote the CLI shows already includes SELAT's ~5% routing markup — it's the final settled charge; nothing is added on top.
- The default Web Search is the wired
selat skill runstep. Agent answer and deep research are separate, agent-run paid calls — each its own spend, so tell the user the cost (from its live quote) and get a yes before every escalation. - Always dry-run first (Step 1 — free, no wallet) for the default step, show the user the real quoted price, and get their OK before any wallet setup or paid run.
- Never ask for, paste, or handle a private key. Wallet auth is the CLI's Circle integration.
Step 0 — get the CLI (free, no account)
selat --version || npm install -g @selat-ai/selat-cliInstalling the CLI creates nothing money-related — no wallet, no account, no keys.
Step 1 — dry run first, before any wallet setup
Do this before creating a wallet or asking the user to fund anything. The dry
run probes the default step's live price and reachability for free — no wallet, no
funds, no account, no selat init:
selat skill install perplexity-search
SELAT_ROUTER_URL=https://router.selat.ai \
selat skill verify ~/.config/selat/skills/perplexity-search(verify takes the installed skill's directory — $XDG_CONFIG_HOME/selat/skills/<name>,
which defaults to the path above. The SELAT_ROUTER_URL prefix is only needed
before selat init has written config.)
This prints the default Search's real quoted price (markup included). Show the user the price and get their OK before wallet setup. If they don't want to proceed, stop here — nothing has been spent or created.
Step 2 — wallet setup (only after the user opts in)
selat init # creates the self-custody Circle Agent Wallet + config
selat fund # deposit USDC into Circle Gateway (user action)
selat doctor # verify wallet, router, and balance are readyselat init is safe to re-run — it detects an existing wallet and asks before
changing anything.
Step 3 — run the default web search
selat skill run perplexity-search \
--query "latest x402 / agentic payments adoption" \
--recency month| Param | Required | Default | What it steers |
query | yes | latest x402 / agentic payments adoption | The web search query. |
recency | no | month | Publication recency filter — one of hour / day / week / month / year. |
The step returns ranked web results with page content and source URLs. Your job: synthesize a concise answer with inline citations to the source URLs, note the recency window, and flag if results are thin or stale. Keep raw JSON out of what you relay — lead with the answer, the sources, and the dollar cost the CLI reported.
Escalating beyond a plain search
When one search pass isn't enough, the installed skill documents two higher-tier
moves — Agent answer (a one-shot synthesized answer) and deep research (an
async sonar-deep-research report, kicked off once then polled for free until it
completes). These are agent-run, not part of the default selat skill run. To
use them:
- Get the live price for the call, tell the user, and get a yes before spending.
- Follow the exact request shapes and steps in the installed skill's own docs —
its
SKILL.md("Escalations") andreferences/endpoints.md— which carry the pinned schemas and the poll loop. Don't guess the request body. - Synthesize the result into a cited brief, same as the default step.
Why this is safe to install
- The skill is a declarative JSON manifest — no executable code. Installing this wrapper only ever writes text.
- Endpoints are https-only and pre-vetted; each publish is gated on a machine-checked live verification receipt.
- Each paid call's spend limit is defined by the underlying SELAT skill and enforced by the runner; escalations are each their own cost-confirmed call, and the runner surfaces the wallet's spending policy at every money moment.
- Funds stay in the user's own wallet. No Perplexity API key, no platform balance, no custodian.
Beyond this skill
SELAT is a general capability layer for paid agent actions. If the user's ask doesn't fit this skill:
selat search "<intent>" # FREE federated discovery + ranking
selat skill list --available # other vetted multi-step skillsDoes it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

