BEAR Transaction Anomaly Analyst

    1

    Analyze synthetic transaction CSVs locally for fan-in/out and repeated pass-through. Get exact evidence IDs, explainable findings, CSV/JSON results, a review report and execution receipts.

    $19

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    BEAR Transaction Anomaly Analyst

    Example session with this skill installed

    Run the bundled synthetic demo with demo/config.json from 2026-01-01T00:00:00Z
    through 2026-01-10T00:00:00Z. Analyze both supported detectors, create a ZIP,
    show the exact flagged entities and why they matched, and report exclusions.

    • Read your context and instructions
    • Compiled the bear transaction anomaly
    • Generated the data export

    Execution completed on the bundled synthetic file: 21 records validated,
    15 eligible, six excluded. Two flags across two account-like entities:

    EntityReason codeExact mechanical evidence
    SYN-FANFAN_IN_OUT_AGGREGATESix distinct sending IDs; 600000 minor units inbound; 420000 outbound; out/in ratio 0.70000000
    SYN-PASSREPEATED_GREEDY_PASS_THROUGHTwo matched rounds; each 2000000 inbound and 1900000 assigned outbound; completion delays 3600 and 7200 seconds

    Supporting IDs

    • SYN-FAN: SYN-FI-1 through SYN-FI-6 and SYN-FO-1.
    • SYN-PASS: SYN-PI-1, SYN-PO-1, SYN-PI-2, SYN-PO-2.

    Excluded: two non-posted rows, one card row, one self-transfer, one unavailable
    at cutoff and one outside the observation. These exclusions remain in the
    validation artifact, not silently removed from the record.

    Delivered files: README.md, results.csv, results.json, evidence.csv,
    analysis-report.md, config.json, input-validation.json, run-receipt.json and
    manifest.json, plus ZIP when requested. Actual local paths are returned after
    execution; this example does not invent a download URL.

    Review the underlying IDs and assignments. The fan-in/out association does not
    prove the incoming funds caused the outflow. The pass-through assignments follow
    the disclosed greedy algorithm, not an economic source-of-funds conclusion.
    No production system was contacted and no real-world risk metric was measured.

    results.csv

    CSV · data export

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Turn a synthetic transaction file into a reproducible, evidence-linked batch analysis. Two inspectable detectors screen account-like IDs for anchored fan-in/out aggregates and repeated chronological pass-through matches.

    The agent runs a self-contained local Python engine with your explicit configuration, run currency, observation start and as-of cutoff. The engine validates every record before analysis, separates excluded records, and gives each result traceable transaction IDs and exact integer-amount features.

    You get:

    • CSV and JSON findings with stable flag IDs and reason codes.
    • Supporting transactions in evidence.csv.
    • A Markdown analysis report with explanations, legitimate lookalikes and limits.
    • Input validation/exclusion details and a receipt with input/config/code hashes, runtime, detector status, settings and actual exit status.
    • Optional ZIP export to a fresh local destination.
    • A complete synthetic demonstration and stored expected results.

    No provider API keys, paid external tools or Python packages are required. Requires an Agensi-connected agent with local shell/file access and Python 3.11+. Runtime verification is recorded in the release status; compatibility with other agent clients is unverified. Without execution capability the analysis is BLOCKED, not a simulated success or a promised downloadable result.

    This early release supports synthetic CSVs only, up to 2000 transactions and 2 MB per input/config file. One run currency (SEK or EUR); mixed-currency input is rejected. It uses explicit configuration, with no calibrated production defaults. The bundled parameters are illustrative and not legal thresholds.

    Fan-in/out is an aggregate screen; outgoing can precede later incoming rows. Pass-through uses deterministic greedy whole-transaction assignment and is not causal funds tracing or balance allocation. A flag can also match legitimate collections, sweeps or settlement behavior. No criminal classification, risk probability, precision/recall or production validation is provided.

    The engine is an adapted MIT-licensed reference algorithm with notices. The product adds the local execution workflow, stricter scope, evidence reporting, failure handling and packaging. Original source is publicly available; no exclusive rights to the upstream detector code are claimed.

    Use the skill for synthetic demonstrations, reproducible mechanics review and analysis training. Customer pilots, measured time savings and market demand have not been established. Your agent/model usage cost is separate.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Finance & Accounting.

    Synthetic-only early release. Max 2000 rows / 2 MB input/config. SEK or EUR in a single-currency run; mixed input rejects the entire file. Fan-in/out is aggregation; pass-through is greedy assignment, not causal tracing. No production validation, calibrated risk scores or measured time savings. External customer pilots pending.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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 today

    Software
    Python 3.11+

    The first release supports synthetic CSV files only. Local execution is required; unavailable clients report BLOCKED instead of simulated results. AI agent with local shell and file access.

    Permissions

    Terminal / Shell

    File Scopes

    user-selected
    synthetic
    CSV/config
    new
    output
    directory

    Local standard-library runtime. No network calls, production connections, dependencies to install or input modifications. Synthetic input acknowledgement required.

    Listed1 day ago

    What's inside

    Frequently Asked Questions

    Analyze synthetic transaction CSVs locally for fan-in/out and repeated pass-through. Get exact evidence IDs, explainable findings, CSV/JSON results, a review report and execution receipts.

    This skill uses the open SKILL.md standard. Use it through an MCP-compatible agent connected to Agensi, and check the compatibility information on this page for confirmed agents and required capabilities.

    Connect your AI agent to Agensi through MCP, then ask it to use BEAR Transaction Anomaly Analyst on your task. Your agent loads the latest approved version each time. Paid skills require a purchase or active subscription. Skill files are not available for download; your licence does not allow you to save, copy or export any part of a skill.

    New in Finance & Accounting