- Home
- Skills
- Research & Analysis
- CSV Export Reconciliation
Works with the AI Agents you already use
CSV Export Reconciliation
Compare two CSV exports by explicit keys and mapped fields. Get row-linked evidence for changed, missing and new records, exact decimal comparisons and duplicate-key blockers.
$9
CSV Export Reconciliation
Example session with this skill installed
Synthetic demo: compare the included left.csv and right.csv using examples/rules.json. Use the declared composite keys and field mappings, preserve the input files, write a new report directory and explain the exceptions and accepted residuals.
- Read your context and instructions
- Compiled the csv export reconciliation
Synthetic demo outcome from the actual local checker: different (exit 1). Three input records per side; one matched pair, one changed pair, one left-only record and one right-only record. Quantity 10.00 versus 10.005 has left-minus-right difference -0.005, accepted within tolerance 0.01 and retained in JSON. Quantity 0 versus 0.00 is numerically equal; missing sides are null. This is instructional synthetic data, not customer evidence or an Agensi end-to-end result.
Connects securely to your tools. The creator never sees your data.
About this skill
Check what changed between two exports, not just whether their totals match.
Use this skill when checking before-and-after migration exports, comparing integration snapshots, or investigating differences between two systems. Your agent confirms the comparison rules, runs the supplied local checker, and explains the evidence without editing your source files.
What it does
- Matches records by an explicit single or composite key, preserving identifiers such as
0007. - Compares selected fields even when column names and row order differ.
- Uses exact decimal arithmetic with a stated absolute tolerance; it does not treat blank numbers as zero.
- Separates changed records, records found on only one side, and nonzero differences accepted within tolerance.
- Blocks duplicate/blank keys and invalid selected values instead of silently dropping records.
- Writes a canonical JSON report with original selected values, source record numbers and input/configuration hashes, plus a CSV review projection.
Example result
In the included synthetic example, each export has three records. The result is one matched record, one changed record, one left-only record and one right-only record. A separate decimal difference of -0.005 is accepted within the declared 0.01 tolerance and remains visible in the report. This is an instructional example, not a customer outcome.
Start with a concrete request
Compare my before and after CSV exports. Help me confirm the record keys, mapped fields, comparable populations and decimal tolerances. Run CSV Export Reconciliation, preserve both inputs, and explain the exceptions with source record numbers.
The package contains the agent workflow, a dependency-free Node.js checker, a configuration reference, a quick start, and runnable synthetic examples.
Requirements and boundaries
Requires Node.js 22+, local file/command access, and a channel that makes the bundled script and references available for execution. The checker itself needs no network calls, API keys or runtime packages; your agent/provider has its own access and data-handling policies. Agensi access is through its agent connection, not a downloadable buyer ZIP.
Each input is limited to 2 MB and 10,000 records, with up to four key components and eight compared fields. Inputs must be UTF-8 with comma, semicolon or tab delimiters. Only selected fields are compared. There is no fuzzy matching, duplicate aggregation, currency conversion, spreadsheet/XLSX parsing, live database integration or source repair. A matching result is not proof of correct business records or a successful migration.
The CSV review file is for convenience; use JSON for exact values. Spreadsheet applications can reinterpret identifiers and formula-like text. The supplied prefixing is not a universal spreadsheet safety guarantee.
Evidence, not guarantees
The implementation has passed local deterministic checks on Windows/Node 24 and Linux/Node 22. These tests are not customer results, an independent benchmark, or certification for every agent. No guaranteed time saving, financial result or universal correctness is promised. Marketplace access, licensing and refunds follow Agensi's applicable terms.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Research & Analysis.
Selected-field, exact-key comparison only. No fuzzy matching, duplicate aggregation, currency conversion, source repair, XLSX parsing or live database integration. Decimal inputs support at most 18 integer and 6 fractional digits, with an explicit nonnegative absolute tolerance. Reports do not certify source truth, comparable populations, correct bookkeeping or a successful migration. No independent customer validation, baseline advantage or Agensi end-to-end execution has been established.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
- Accounts
Agensi- Software
Node.js 22+Terminal access
You provide two local UTF-8 CSV exports and an approved comparison contract. The agent must be allowed to run a temporary script.
Permissions
Allowed Hosts
File Scopes
The checker does not make network requests, install packages or modify source data. Agensi serves skill text through its connection. Agent/model-provider access and data-handling policies apply separately. No input or report upload is needed.