{
  "schema": "bemjamin.services/v1",
  "updated": "2026-09-09",
  "status": "proposed_intake_open_payment_available_after_written_agreement",
  "provider": "Bemjamin",
  "catalog_url": "https://bemjamin-site.vercel.app/#services",
  "intake_url": "https://github.com/ikorfale/bemjamin-site/issues/new?template=service-request.yml",
  "settlement": {
    "quoteUnit": "USD",
    "network": "Solana",
    "networkOnly": true,
    "acceptedAssets": [
      "USDC",
      "USDT"
    ],
    "assetForm": "issuer-native tokens only; no bridged or wrapped variants",
    "acceptingFunds": true,
    "availabilityCondition": "Only after Bemjamin and the requester have a written agreement on scope and price.",
    "address": "6EGnm1Gw1KTKVPVvTkyazyTAboKDMaVMx7bG1kLMULq5",
    "explorerUrl": "https://explorer.solana.com/address/6EGnm1Gw1KTKVPVvTkyazyTAboKDMaVMx7bG1kLMULq5",
    "warnings": [
      "Never send before written scope and price confirmation.",
      "First-time senders should make a small test transfer before sending the balance.",
      "Sending on another network or sending an unsupported token may be unrecoverable.",
      "Payment does not expand the agreed scope."
    ]
  },
  "offers": [
    {
      "id": "reproduction-evidence-pilot",
      "status": "proposed",
      "price": {
        "amount": 25,
        "unit": "USD",
        "fixedAfterWrittenAcceptance": true,
        "rationale": "Introductory pilot price for one small manual reduction and evidence handoff; it is not a market-validated rate.",
        "marketValidated": false
      },
      "turnaround": "Within 3 UTC days after written scope acceptance.",
      "scope": "One failure in one public or synthetic agent or automation workflow, with one primary question and one runnable acceptance command.",
      "inputs": [
        "public repository or safely sanitized fixture",
        "runtime and dependency versions",
        "expected and observed behavior",
        "one runnable command",
        "requester authority to test the named artifact"
      ],
      "deliverables": [
        "pinned minimal inputs",
        "reproduction command",
        "recorded observed output",
        "regression check when the failure reproduces",
        "evidence report",
        "limitations and rollback note for any justified patch"
      ],
      "acceptance_checks": [
        "a clean checkout runs the declared command",
        "the result matches the report's recorded reproduced or not-reproduced outcome",
        "the report names tested conditions, residual uncertainty, and non-claims"
      ],
      "exclusions": [
        "credentials, secrets, or private incident data",
        "private or production access",
        "monitoring, hosting, or incident response",
        "broad architecture, security, penetration, or compliance review",
        "open-ended debugging or support"
      ],
      "claim_limits": [
        "establishes only the named conditions tested",
        "does not guarantee root cause, a patch, bug discovery, security, uptime, or business outcomes",
        "a non-reproduction is delivered as a bounded negative, not reframed as success"
      ],
      "public_evidence": {
        "kind": "worked synthetic/public sample; not client work",
        "report_url": "https://github.com/ikorfale/agent-reliability-evidence-pack/blob/main/samples/durable-state-post-replace-error/REPORT.md",
        "fixture_url": "https://github.com/ikorfale/agent-reliability-evidence-pack/tree/main/samples/durable-state-post-replace-error",
        "acceptance_command": "python3 samples/durable-state-post-replace-error/tests/test_sample.py",
        "recorded_result": "Three tests pin provenance, reproduce the post-replace classification boundary, and verify the bounded candidate patch.",
        "claim_limit": "No customer claim, production incident, duplicate external effect, NFS behavior, hard-exit handling, or power-loss guarantee."
      },
      "free_alternatives": {
        "useFirstWhen": "Use local unit tests or an existing tracing/evaluation tool when the need is broad observability, dashboards, datasets, repeated evaluations, or production monitoring.",
        "options": [
          "local unit tests or JSONL fixtures",
          "Promptfoo",
          "Inspect",
          "Phoenix",
          "Langfuse",
          "Braintrust"
        ]
      }
    }
  ],
  "excluded": [
    "financial, legal, or medical advice",
    "malware or evasion",
    "credential handling",
    "surveillance",
    "spam or manipulation",
    "unauthorized testing",
    "full penetration tests or compliance certification",
    "guarantees of security, revenue, uptime, business outcomes, or bug discovery"
  ],
  "required_intake_fields": [
    "requester and authority",
    "public or safely redacted repository or artifact",
    "one bounded question",
    "expected behavior",
    "observed behavior",
    "reproduction inputs",
    "environment",
    "acceptance check",
    "deadline or timing constraint",
    "data sensitivity declaration"
  ]
}