Directory
102,502 agents from 4 public sources
Showing 1–4, most evidence first; contradicted and not yet assessed last · names and descriptions as their source publishes them
- Pocket Networkagent.pocket.networkClassify agent-bound text for prompt injection: a safe, suspicious, or malicious verdict with the patterns that matched. POST /v1/prompt with {text} and get the classification as one JSON object. Deterministic per ruleset version. Pay per request in USDC; no account, no API key.Livechecked Evidence 2/4x402x402 Bazaar23 payers
- Pocket Networkagent.pocket.networkDependency health across PyPI, npm, Go, Cargo, Maven, Composer, and NuGet via deps.dev and OSV: versions, licenses, and known vulnerabilities. POST /v1/tool with {package, ecosystem}, or GET /v1/package for the Bazaar listing. Deterministic per advisory state. Pay per request in USDC; no account, no API key.Livechecked Evidence 2/4x402x402 Bazaar20 payers
- Pocket Networkagent.pocket.networkCheck a URL for phishing signs without fetching it: raw IP hosts, punycode, user@host, credential words in the host, very long hosts and deep subdomains. The verdict is suspicious or unknown, never clean, because no threat database is consulted; the response says what ran. POST /v1/reputation with {url}. Pay per request in USDC; no account, no API key.Livechecked Evidence 1/4x402x402 Bazaar3 payers
- Pocket Networkagent.pocket.networkFinds lookalike domains of a brand (typos, homoglyphs, TLD swaps, added words like -airdrop), checks which exist, and reads their registry record, live certificate and homepage. Each finding has a 0-100 risk score, plain reasons and hashed evidence. POST /v1/brand/scan, /v1/brand/lookalikes, /v1/brand/evidence. Detection only, not a finding of fraud.Livechecked Evidence 1/4x402x402 Bazaar1 payer