Quietstore / x402 on Base You watch · agents pay

No accounts · a few cents of USDC

Point your agent here. Read the receipts.

Agents on Bazaar already pay a cent or two for live jobs: weather, prices, “audit this URL,” “is this npm package safe.” Quietstore is that shape — fetch your public site after a deploy, screen the lockfile against npm and CVEs — plus a report you can read. No signup. The badge goes grey in a week if the agent stops.

Give the agent /llms.txt (or /v1/catalog). You watch the origin report and the lockfile report. Bazaar readiness is free if you sell x402. Live origin and lockfile doctors are $0.02 a call.

01 / People

You point. The agent pays. You read.

No signup. No API key. Paste the store URL into the agent (or tell it to read /llms.txt). It already speaks x402.

Two cents after a deploy. Two cents after a lockfile change. One cent if you sell x402 and Bazaar is broken. The agent spends USDC from its wallet; you never hold a seller key here.

  1. Tell the agent this host. First fetch: /llms.txt or /v1/catalog.
  2. After it deploys, it POSTs your public URL. You open the origin report.
  3. After it installs packages, it POSTs package.json. You open the lockfile report.
  4. Put the badge in the README. It goes stale in seven days on purpose — silence is visible.
02 / Tools

Live lookups first

Two cents to fetch a URL or screen npm. One cent to fix Bazaar. The last three are optional extras.

Prove the agent's deploy claims against the live origin

$0.02 / call

Fetch a live https URL and score whether the deploy is actually there. POST {"url":"https://your.site","min_score":80} ($0.02). Default x402-seller profile: HTTPS, HSTS, llms.txt, catalog, OpenAPI, MCP, health. Optional claims[] or profile:"web". Returns 0–100, evidence, do_this_first, ci_failed. Humans: GET /v1/tools/claim-check/report?host=your.site. Badge goes stale after 7 days.

POST /v1/tools/claim-check

Lockfile doctor: live npm + OSV punch list

$0.02 / call

Screen declared npm packages before you install them. POST package.json (optional lockfile) ($0.02). After payment we look up each name on the npm registry and OSV: missing/hallucinated packages, lookalikes, brand-new publishes, deprecations, known CVEs. Returns 0–100, do_this_first, ci_failed. Caps at 60 packages; never fetches buyer URLs. Report: GET /v1/tools/dep-doctor/report?project=your-app.

POST /v1/tools/dep-doctor

Scan a live URL → 0–100 hygiene score + ranked fixes

$0.02 / call

Audit a public https URL. POST {"url":"https://your.site","min_score":80} ($0.02). After payment we fetch the origin (headers, HSTS, CSP, TLS, cookies, DNS/email, security.txt) and return a 0–100 score, do_this_first, and ci_failed for CI. Free preview is snapshot-only {headers}. Badge: GET /v1/tools/site-hygiene-score/badge?host=your.site.

POST /v1/tools/site-hygiene-score

Will my x402 endpoint show up on Bazaar?

$0.01 / call

Diagnose why an x402 seller is invisible on Coinbase Bazaar. POST your 402 PaymentRequired JSON. Free preview: pass/fail + which checks fail. Paid $0.01: every finding with the exact fix (extension shape, AJV schema, description limits). Run before each x402 deploy.

POST /v1/tools/x402-bazaar-audit

Turn a react-doctor JSON report into a ranked 0–100 punch list

$0.02 / call

Run `npx -y react-doctor@latest --json .` locally, POST the JSON here. Free preview at POST /v1/tools/react-doctor-report/preview. Paid $0.02 returns do_this_first with help text and file:line samples, plus min_score/ci_failed. The Worker never scans your repo. Uses the report's official summary.score when present (react-doctor 0.9+ score API); otherwise estimates with unique-rule penalties.

POST /v1/tools/react-doctor-report

Lint an agent handoff before you switch owners

$0.005 / call

POST a handoff note → score plus what's missing (verification evidence, constraints, ownership, next steps) before you switch owners. $0.005/call. Guide: pack demo-notes-v1 ($0.01).

POST /v1/tools/handoff-lint

Will this shell command blow my agent token budget?

$0.005 / call

POST a shell command → token-budget risk score, whether to wrap it with RTK, and a safer form. Never executes the command. $0.005/call. Full RTK workflow: pack rtk-output-discipline-v1 ($0.05).

POST /v1/tools/rtk-command-score

Should I launch more coding agents right now?

$0.01 / call

POST your capacity metrics (running vs requested agents, host load, cores, ready work) → admit / reduce / refuse with reasons. $0.01/call; you send the numbers, the store never touches your host. Full recipe: pack beads-agent-mail-recipe-v1 ($0.10).

POST /v1/tools/swarm-admission
03 / Packs

Deep guides (one-shot)

Coordinate Multiple Coding Agents Without Collisions

$0.10

Operational recipe for multi-agent work on one repo: claim tasks, reserve files, avoid stomping, commit clean state, and hand off safely. Companion tools: POST /v1/tools/swarm-admission ($0.01) before launch and POST /v1/tools/handoff-lint ($0.005) on each handoff.

beads-agent-mail-recipe-v1

Fresh-Eyes Review Checklist for Agent Work

$0.005

Smallest paid pack ($0.005): a machine-readable merge checklist and a one-request smoke purchase on this store. After you verify the rail, browse GET /v1/tools for free Bazaar-audit preview and metered loops.

demo-extras-v1

Hand Work Between Agents Without Losing the Plot

$0.01

Compact handoff standard so the next agent keeps constraints, verification evidence, ownership, and next steps. For live scoring of any handoff text, use POST /v1/tools/handoff-lint ($0.005 per call).

demo-notes-v1

Prove an x402 Purchase Works End-to-End

$0.01

A five-minute drill for standing up an x402 seller: catalog → 402 → pay → body → replay → ledger, each step with a check. Run the free x402-bazaar-audit preview alongside it so agents can actually discover you on Bazaar.

dod-drill-v1

Stop Wasting Tokens on Noisy Shell Output

$0.05

Practical RTK workflow for coding agents: when to compress command output, keep exit codes, and avoid log floods that burn the context window. Score a command first with POST /v1/tools/rtk-command-score ($0.005) — does not execute the command.

rtk-output-discipline-v1

Live URL Hygiene Score Agents Re-run After Every Deploy

$0.05

Live {url} scan loop for Quietstore's site-hygiene-score tool: pay $0.02, we fetch the public origin, return a 0–100 score plus do_this_first fixes, min_score CI gate, and delta vs last scan. Snapshot preview stays free. SSRF-hardened; stored replay (no second fetch).

site-hygiene-score-v1

Sell to Agents with x402 — No Double Charges

$0.25

A production playbook for charging AI agents over HTTP 402 without double-billing: safe retries, ambiguous-settle handling, cache rules, ledger idempotency, and buyer recovery messages — the exact patterns behind this store, live on Base USDC. Pairs with the free x402-bazaar-audit preview to check your own endpoint.

x402-retry-safe-seller-v1
04 / API

For agents

Machine entry: /v1/catalog (pitch, start_here), /v1/tools, /llms.txt, /openapi.json, /mcp. Free Bazaar check: POST /v1/tools/x402-bazaar-audit/preview.

If a paid request fails mid-flight, retry with the SAME X-PAYMENT header: an already-settled payment is re-served without a second charge. Signing a NEW payment authorization is a NEW purchase. Metered tools: same payment + same JSON body re-serves; a new body needs a new payment.

  1. Orient.
    curl -sS https://<store-host>/v1/catalog
  2. Free Bazaar preview (no payment).
    curl -sS -X POST https://<store-host>/v1/tools/x402-bazaar-audit/preview \
      -H 'content-type: application/json' \
      -d @payment-required.json
  3. Paid live check (402 until paid).
    curl -i -X POST https://<store-host>/v1/tools/claim-check \
      -H 'content-type: application/json' \
      -d '{"url":"https://your.site","min_score":80}'
  4. Pack example.
    curl -i https://<store-host>/v1/packs/demo-notes-v1

Verify what you bought

content_sha256 is SHA-256 over the UTF-8 bytes of the returned body. Check byte_len first, then compare the digest with both the paid response and catalog:

jq -j '.body' response.json | shasum -a 256

jq -j is intentional; -r adds a newline and changes the hash.

Terms

Purchased content is licensed for the buyer's own use (human or agent); no resale or redistribution.