generated: '2026-09-19'
method: searched
source: https://github.com/OrderofChaos33/SINCOR2/blob/main/docs/A2A_EXTERNAL_WIRE.md
corroboration:
  - https://getsincor.com/api/a2a/quote?skill_id=lead-enrichment      # is_free true, free_quota_remaining 5 (probed)
  - https://github.com/OrderofChaos33/SINCOR2/blob/main/examples/a2a_external_caller.py   # --simulate flag; default base URL local
  - https://github.com/OrderofChaos33/SINCOR2/blob/main/README.md      # local run on :8080; curl localhost:8080/.well-known/agent-card.json
  - https://github.com/OrderofChaos33/SINCOR2/blob/main/docs/LOOP_SETTLEMENT.md   # Base Sepolia testnet loop
  - https://github.com/OrderofChaos33/SINCOR2/blob/main/docs/A2A_PRODUCTION_CHECKLIST.md
checked: '2026-09-19'
summary: >-
  There is no separate sandbox host or test-mode key. What SINCOR publishes instead is (1) a FREE, ANONYMOUS
  path on production — six skills allow 5 calls per caller_id, and the provider's own probe log records those
  calls executing with simulation_mode set and never touching the realized ledger; (2) a quote endpoint that
  returns price, fee split and the per-skill JSON Schemas without side effects; (3) simulated settlement:
  transaction hashes prefixed 0xSIMULATED… are recognised and "never hit the realized ledger"; (4) an
  open-source runtime you can run locally on port 8080 with the same card and endpoints; and (5) a Base
  Sepolia testnet settlement loop the team has documented. No test card numbers or hosted test tokens exist —
  payment is an on-chain ERC-20 transfer and there is no test-mode key prefix.
modes:
  production: https://getsincor.com/api/a2a
  local: 'http://localhost:8080 — `git clone https://github.com/OrderofChaos33/SINCOR2.git && pip install -r requirements.txt && python run.py` (README Quickstart)'
  testnet: 'Base Sepolia (docs/LOOP_SETTLEMENT.md) — testnet USDC escrow, ERC-7579 session-key payout; a console tape, not a public sandbox endpoint'
key_prefixes: null   # no API keys are issued (see authentication/)
free_quota:
  calls: 5
  scope: per caller_id per skill
  skills: [lead-enrichment, competitor-intel, outreach-sequence, market-forecast, content-blog]
  how: 'include caller_id in the tasks/send / message/send request; no txHash needed (quote note, verbatim: "FREE — include caller_id in your tasks/send request (no txHash needed).")'
  behaviour: 'metadata.caller_id=anonymous, free_call=true, simulation_mode (provider probe log 2026-09-10)'
simulation:
  tx_hash_prefix: '0xSIMULATED…'
  effect: 'Simulated (0xSIMULATED…) and free-quota tasks never hit the realized ledger (docs/A2A_EXTERNAL_WIRE.md); "Simulated and free-quota paths write zero realized events" (docs/A2A_PRODUCTION_CHECKLIST.md)'
  client_flag: 'python examples/a2a_external_caller.py --simulate — "never posts real payment intents; prints the full flow"'
dry_run:
  endpoint: https://getsincor.com/api/a2a/quote?skill_id=<skill>
  returns: [axm_price_wei, axm_price_display, is_free, free_quota_remaining, pay_to, platform_fee_bps, treasury_fee_split, estimated_latency_seconds, input_schema, output_schema]
  side_effects: none
test_values:
  cards: null
  bank_accounts: null
  hosted_tokens: null
  test_clocks: null
  note: 'Never invented; the provider publishes none. Payment is a wallet transfer on Base.'
fixtures:
  - 'examples/agent_cards/*.json — reference agent cards (healthcare_credentialing, trading_bot, compliance_agent, toa_temporal_optimization, LiquidityManager)'
  - 'examples/a2a_loop_demo.py — discovery → task submission → settlement demo against a local instance using the lead-generation vertical'
  - 'examples/workflows/multi_agent_pipeline.json'
caveats:
  - 'The free path runs on PRODUCTION and shares the production task store; it is a quota, not an isolated environment.'
  - 'Free quota is bound to a self-asserted caller_id, so exhaustion is per string, not per account.'