Client Report QA

    2

    Check a pasted KPI table, flag misleading comparisons, and draft a short source-linked client update.

    Free

    6 uses5.0 (1 review)

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Client Report QA

    Example session with this skill installed

    Self-initiated synthetic demonstration, not client work or a business performance result. Output comes from a local independent AI test; a human must review it.

    Please check my monthly client report and write the short update. This is a synthetic demo for Client A. Compare July 1-31 and August 1-31, 2026, UTC, USD. Both columns use the same ads account, complete periods, definitions and attribution. R1 spend: 1000 then 1200 USD, excluding fees. R2 reported leads: 40 then 60, same conversion action; not unique people. N1: We changed the headline August 10 but ran no experiment. Source is the pasted synthetic ads export. Use only these numbers.

    • Read your context and instructions
    • Compiled the client report qa

    Review Status

    READY FOR HUMAN REVIEW. The supplied context supports the comparisons in the draft, and all derived arithmetic was tool-checked. This status is not approval to send and does not independently verify the pasted export.

    Local labels below identify pasted material, not independently verified sources. Labels are scoped to this request.

    Internal Checks

    SourceSupplied values or contextFormula or comparison exclusionCalculation statusUnresolved issues
    C1Synthetic Client A; July 1-31 and August 1-31, 2026; UTC; USD. Same ads account, complete periods, definitions and attribution. Source: pasted synthetic ads export.Like-for-like comparison supported by supplied context.Context reviewed; not independently verified.None blocking this draft.
    R1Spend excluding fees: USD 1,000.00 -> USD 1,200.00.Change = 1200 - 1000 = USD +200.00. Relative change = (1200 - 1000) / 1000 * 100 = +20.0%.TOOL-CHECKED.Upstream source accuracy not verified.
    R2Reported leads: 40 -> 60; same conversion action; not unique people.Change = 60 - 40 = +20. Relative change = (60 - 40) / 40 * 100 = +50.0%.TOOL-CHECKED.Counts must not be represented as unique people or customers.
    R1 + R2Matching period spend and reported leads.Prior cost per reported lead = 1000 / 40 = USD 25.00. Current = 1200 / 60 = USD 20.00. Change = 20 - 25 = USD -5.00. Relative change = (20 - 25) / 25 * 100 = -20.0%.TOOL-CHECKED.This is cost per reported lead, not customer acquisition cost or a profitability measure.
    N1Headline changed August 10; no experiment was run.Work performed and observed outcomes are separate. No causal attribution to the headline edit.No arithmetic required.The edit's causal effect is not established.

    Rounding: displayed currency uses two decimals; percentages use one decimal. Calculations retained unrounded values until display.

    Client Draft

    Client A's July 1-31 and August 1-31, 2026 reporting periods are complete and comparable, using UTC and USD. C1

    Spend excluding fees increased from $1,000.00 to $1,200.00 (+20.0%), while reported leads increased from 40 to 60 (+50.0%). R1 R2 Cost per reported lead decreased from $25.00 to $20.00 (-20.0%). R1 R2 These leads are not unique people. R2

    The headline changed on August 10, but no experiment was run, so these results do not establish that the edit caused the change. N1

    Proposed next action: design a controlled headline test for the next reporting period.

    Next Input

    None; human approval still required.

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

    About this skill

    Review the numbers before writing the story. Client Report QA helps freelancers and small marketing teams turn a redacted KPI table or monthly notes into a compact client draft with an internal check trail.

    PASTE WHAT YOU ALREADY HAVE

    A small table or plain text is enough to start. Add the client alias, reporting periods, units and source context when you know them. The skill asks at most three necessary questions and works with the usable portion immediately.

    WHAT YOU GET

    • A READY FOR HUMAN REVIEW or HOLD status with the specific reason
    • A source-labeled check table and missing-data or conflict flags
    • A client update of up to 180 words
    • One bounded next action and only the questions that matter

    USEFUL CHECKS

    Partial periods are not treated as monthly growth. Blank values are not zero. Conflicting exports are not silently reconciled. Percentage points and percent change are kept separate. A campaign edit is not automatically credited for an outcome.

    BOUNDARIES

    This is a text-based preparation workflow, not an analytics connector, dashboard or autonomous reporting service. Derived arithmetic needs a trusted host calculation tool; without one, derived values are marked unverified and withheld from the client draft. Source accuracy and final approval remain human responsibilities. No account connection, network access or API key is needed. Only use redacted material you are permitted to process. AI-host charges and data policies remain separate.

    The demonstration uses synthetic data, not client results. No time-saving, revenue or accuracy guarantee is made.

    Does not fetch or independently verify analytics, reconcile unknown conflicts, infer causality, or send reports. Checked arithmetic requires a trusted host calculation tool; without it, derived numbers are withheld from the client draft. Outputs require human review. Tested on synthetic examples, not a cross-model benchmark or client study.

    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 any payment.

    Reviews

    5.0
    1 review
    5
    1
    4
    0
    3
    0
    2
    0
    1
    0

    This skill has been used 6 times.

    Trust & safety

    Security scan

    Security scan passed 2 months ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed2 months ago

    What's inside

    Frequently Asked Questions

    No. Paste a redacted table or notes you are allowed to process. The skill does not log in, fetch analytics or request an API key.

    The skill marks affected comparisons as unresolved, produces the usable portion, and asks up to three relevant questions. It does not pick an authoritative number on your behalf.

    No. Even READY FOR HUMAN REVIEW still needs your approval. A calculation tool checks arithmetic, not whether the upstream source is true. Without that tool, derived numbers are excluded from the client draft.

    New in Marketing & SEO

    Free