ContentSweep / Supplier Feeds

Reproducible synthetic demonstration

Reuse a supplier mapping and stop structural drift

Process two compatible deliveries with one explicitly reviewed profile, preserving identifiers and decimals. Stop a changed header before another evaluation.

2 deliveriesOne approved profile reused
000123 / 19.90Identifiers and decimals preserved
Review requiredStructural drift stops the next start

These are actual recorded results from owned synthetic fixtures. Native Gemini workflow verified. The recorded hosted runs used internal noncash allowances and are excluded from customer revenue. This page runs no jobs and needs no account. Live hosted execution requires your own private key and authorization.

Run with your agentDownload the example

Inspect the actual result

Successful results and deliberate stop cases are both part of the demonstration. Generated IDs, times and machine paths can differ; findings, business values, coverage and accounting must agree.

Native Gemini workflow verified. Product runtime: ai.contentsweep/supplier-feeds 1.1.0.

Complete recorded result (JSON) · Native agent evidence · Fixtures and checksums

Read the result evidence
{
  "deliveries": [
    {
      "case": "first",
      "output_sha256": "32379aff610cb7d0a69b5ab3ff77213d5a4c97588a6d5f112336813c7bf285e3",
      "report": {
        "error_count": 0,
        "errors": [
          
        ],
        "errors_truncated": false,
        "evaluated_rows": 2,
        "full_file_validation": true,
        "quality_gate": {
          "coverage": {
            "policy": "analyzed",
            "profile": "analyzed"
          },
          "findings": [
            
          ],
          "generated_at": "2026-11-13T15:05:17.522246Z",
          "input_fingerprint": null,
          "metadata": {
            "contract": {
              "rules": [
                {
                  "column": "sku",
                  "id": "unique-sku",
                  "type": "unique"
                }
              ]
            },
            "profile": {
              "columns": {
                "currency": {
                  "distinct_count": 1,
                  "null_count": 0
                },
                "price": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "sku": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "stock": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "title": {
                  "distinct_count": 2,
                  "null_count": 0
                }
              },
              "row_count": 2
            }
          },
          "product": "data-quality-gate",
          "schema_version": "1.0",
          "verdict": "continue"
        },
        "row_count": 2,
        "semantic_guarantee": "approved mapping and explicitly configured checks only",
        "verdict": "passed"
      },
      "retry_same_job": true,
      "rows": [
        {
          "currency": "EUR",
          "price": "19.90",
          "sku": "000123",
          "stock": 12,
          "title": "Desk lamp"
        },
        {
          "currency": "EUR",
          "price": "24.50",
          "sku": "000124",
          "stock": 8,
          "title": "Reading light"
        }
      ]
    },
    {
      "case": "later",
      "output_sha256": "c7e2b3699439588abe299b763b983937a6aad95f9f5f6c4e89d3c93f664d6601",
      "report": {
        "error_count": 0,
        "errors": [
          
        ],
        "errors_truncated": false,
        "evaluated_rows": 2,
        "full_file_validation": true,
        "quality_gate": {
          "coverage": {
            "policy": "analyzed",
            "profile": "analyzed"
          },
          "findings": [
            
          ],
          "generated_at": "2026-11-13T15:05:24.951462Z",
          "input_fingerprint": null,
          "metadata": {
            "contract": {
              "rules": [
                {
                  "column": "sku",
                  "id": "unique-sku",
                  "type": "unique"
                }
              ]
            },
            "profile": {
              "columns": {
                "currency": {
                  "distinct_count": 1,
                  "null_count": 0
                },
                "price": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "sku": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "stock": {
                  "distinct_count": 2,
                  "null_count": 0
                },
                "title": {
                  "distinct_count": 2,
                  "null_count": 0
                }
              },
              "row_count": 2
            }
          },
          "product": "data-quality-gate",
          "schema_version": "1.0",
          "verdict": "continue"
        },
        "row_count": 2,
        "semantic_guarantee": "approved mapping and explicitly configured checks only",
        "verdict": "passed"
      },
      "retry_same_job": true,
      "rows": [
        {
          "currency": "EUR",
          "price": "19.90",
          "sku": "000123",
          "stock": 13,
          "title": "Desk lamp"
        },
        {
          "currency": "EUR",
          "price": "24.50",
          "sku": "000124",
          "stock": 8,
          "title": "Reading light"
        }
      ]
    }
  ],
  "destination": "local preview only",
  "drift": {
    "billable": false,
    "reason": "source structure changed",
    "source_structure": {
      "fields": [
        "available",
        "currency",
        "product_name",
        "unit_price",
        "vendor_sku"
      ],
      "fingerprint": "321dc357bc41eb2eac3906f77b999cf7e25ddf219e94d33c9d7a5755077a4fdc",
      "types": {
        "available": [
          "string"
        ],
        "currency": [
          "string"
        ],
        "product_name": [
          "string"
        ],
        "unit_price": [
          "string"
        ],
        "vendor_sku": [
          "string"
        ]
      }
    },
    "status": "review_required"
  },
  "maximum_units": 2,
  "same_approved_profile": true
}

Download SHA256: 9eac8b2eb10eaa45659438b1c32769ed0afc32d4e9d220bc54c8b19d4d2fa8f5 · Checksum file

Process two compatible deliveries with one explicitly reviewed profile, preserving identifiers and decimals. Stop a changed header before another evaluation.

Owned synthetic example. Recorded results are clearly separated from live execution. Installing a skill does not install its runtime or authorize spending or uploads.

Inspect the evidence

Read manifest.json, expected.json, and the fixtures before running. The walkthrough covers first, later, drift. Results come from actual product MCP calls, not simulated responses. The free public page and downloads require no sign-in.

Reproduce with your agent

Read the existing supplier-feed-import skill and its companion setup/tool references. Then follow AGENT_PROMPT.md. Create a dedicated Python 3.12+ environment and install mcp>=2.3,<3. Configure CONTENTSWEEP_SUPPLIER_FEEDS_API_KEY privately in your environment; never paste its value into prompts, command arguments, URLs or source control.

These commands are qualified on macOS and Linux. Follow client setup for runtime installation. Use that environment’s Python to run the demo.

For deterministic verification from this repository root:

python demo/run.py --mode hosted --max-units 2 --output demo-results

The hosted allowance explicitly authorizes at most 2 evaluations of these named synthetic cases. Review the supplied field decisions before approving them. It grants no permission for customer files, purchases, destination writes, source edits or cloud actions. Hosted keys are your own; local analysis needs no hosted key.

Reuse the same output directory and allowance after an uncertain response. resume.private.json preserves quote/key/job identities before dispatch. Keep it private. Never start a new run merely because polling timed out. Changed customer data needs its own review and authorization. A completed rejection, block or warning can be billable as documented for this product.

Interpret the result

Inspect result.json for assertions, complete reports and actual tool discovery. Expected stop cases are part of a passing demo. A deterministic runner is not a native agent completion. Machine-generated IDs, timestamps and local paths may differ; business values, coverage and accounting must agree.

Flat CSV/JSONL, approved mappings, 10 MiB, 50,000 records, 100 fields. No automatic destination writes.

Price and next step

Five trials per verified organization; $20/10 evaluations or $99/month/100. Completed rejections consume an evaluation; exact retries do not.

Read the full walkthrough and setup · Product and current terms

Keep local/free options separate from optional hosted purchases. This example performs no purchase. Reports are bounded evidence, not runtime certification or proof of customer outcomes.

Preserve original bytes in a native agent session

Some clients reconstruct file contents or rewrite object keys. Use this reviewed helper through the agent’s native shell tool to upload the exact checksum-bound fixtures, then continue with the product’s native MCP review, approval, quote, evaluation, report and retry tools:

python demo/transfer.py --confirm-upload --output demo-uploads/uploads.private.json

It uploads only the bundled synthetic files and prints owned upload IDs. It performs no approval, evaluation, purchase or destination write. Keep its identity file private and reuse it after an interruption. This mixed file-transfer/MCP workflow is disclosed in the native evidence.

Recurring delivery workflow

The bundled fixtures/recurring-import-reviewed.n8n.json is the existing reviewed recurring workflow. Read fixtures/N8N_GUIDE.md for approval, compatible delivery, receipt verification, drift stop and retry recovery. Importing it does not activate a schedule. Inspect first accepted output and later accepted output; their complete byte hashes match the recorded receipts.

Copy the agent task · Read this walkthrough as Markdown · Client setup and recovery