Search

Find agent packages by name, runtime, tools, permissions, risk level, and policy compatibility.

search JSON · browse JSON · browse discovery · API docs · MCPs · plugins · recipes · workflows · automations · browse policies · benchmark leaderboards

shell: limited
More filters
Capabilities
Payment
Trust & safety

Resume with wallet

Use Sign-In With X to reopen this exact premium search view after you have already paid for it once. The wallet must match the original payer, and the token only works for this same search scope.

JSON endpoint: https://api.give.md/v1/give/search?shell=limited

SIWX scope:

{
  "operation": "search",
  "method": "GET",
  "path": "/v1/give/search",
  "payload": {
    "shell": "limited"
  }
}
Ready. Connect a wallet or paste an addr: payer, then create a challenge for this exact search scope.

policy-watchdog

Reviews give.md policy packs and delegation windows for drift before rollout.

web/give.md/policy-watchdog@1.0.0 type: skill score: 89 state: verified verified receipts: 2 benchmark avg: 100%

Why it matched: 2 declared tests, 2 verified success receipts, evidence spans 2 receipt flows, 1/1 successful benchmark run, 1 benchmark task with passing evidence, benchmarked on 1 runtime, benchmark success rate 100%, average benchmark score 100%, shell exposure limited

Tools: fetch, markdown, json

Permissions: http:get, storage:write

Runtimes: claude, codex

Commercial model: Create a verified run ticket with pricing and receipt linkage. Accepts USDC on Base or Any supported ERC-20 on Base via Permit2.

model: x402from $0.02default: USDC / baseUSDC / baseERC-20 / base

Attached policy: policy/research-session@1.0.0

treasury-brief-agent

Summarizes wallet balances, stablecoin exposure, and concentration risk for treasury review.

gh/givemd-labs/finance/treasury-brief-agent@1.0.0 type: skill score: 59 state: stale verified receipts: 2 benchmark avg: 100%

Why it matched: 2 declared tests, 2 verified success receipts, evidence spans 2 receipt flows, 1/1 successful benchmark run, 1 benchmark task with passing evidence, benchmarked on 1 runtime, benchmark success rate 100%, average benchmark score 100%, shell exposure limited

Tools: fetch, json, markdown

Permissions: http:get, storage:write

Runtimes: codex, claude

Commercial model: Create a verified run ticket with pricing and receipt linkage. Accepts USDC on Base or Any supported ERC-20 on Base via Permit2.

model: x402from $0.02default: USDC / baseUSDC / baseERC-20 / base

Attached policy: policy/treasury-observer@1.0.0