More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Observability Reproduction Designer

    1

    Turn intermittent incidents into safe telemetry and falsifiable tests.

    $19.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Observability Reproduction Designer

    Observability Reproduction Designer

    Example session with this skill installed

    Design a reproduction plan for intermittent checkout timeouts after payment authorization. We have UTC millisecond timestamps, gateway status and latency, region, payment type, cart-size bucket, release ID, and a generic timeout error. Gateway events cannot be joined to worker attempts. The worker retry ordinal and order-state transition are missing. Raw customer and payment fields must never be logged. Create competing hypotheses, the minimum telemetry delta, a controlled protocol, and stop rules.

    • Read your context and instructions
    • Compiled the observability reproduction designer

    Verdict — INSTRUMENTATION-REQUIRED

    The symptom is precise enough to form competing hypotheses, but current evidence cannot join a late gateway response to a specific worker attempt or order-state transition. One bounded event is required before a responsible reproduction attempt.

    Symptom contract

    Intermittent checkout timeout observed after payment authorization. Known-good comparator: successful checkouts with the same region, payment type, cart-size bucket, and release ID. Success criterion for reproduction: controlled execution reaches authorization, produces overlapping or delayed worker behavior, and ends in the same timeout category and terminal order state.

    Evidence map

    DomainStatusEvidence
    Correlationgapgateway and worker identifiers cannot be joined
    TimeavailableUTC millisecond timestamps
    Cohortgapregion available; tenant grouping absent
    Inputavailablebounded cart size and payment type
    Application stategaporder transition not emitted
    Dependency stateavailablegateway status and latency
    Retrygapattempt ordinal missing
    Erroravailabletimeout category
    Versionavailablerelease ID
    Privacyavailableraw payment data excluded

    Competing hypotheses

    H1 — A worker retry overlaps a late gateway response. Support: failures follow authorization. Contradiction: some samples show only one worker log. Falsifier: correlated failures contain no overlapping attempt ordinals.

    H2 — One region experiences dependency latency spikes. Support: weak regional clustering. Contradiction: the failure sample is small. Falsifier: matched success and failure cohorts show equivalent gateway-latency distributions.

    Minimum instrumentation

    Emit checkout_transition immediately before and after the gateway boundary with hashed_correlation_id, bounded state_from and state_to, attempt_ordinal, region, and release_id. Capture 100% of failures and a 1% success comparator for 14 days. Exclude raw customer, cart, token, and payment fields. Bound all dimensions to enumerations or an approved hash. Remove the event after the hypothesis verdict and retention review.

    Controlled reproduction

    Freeze a gateway stub at selected late-response intervals, replay one bounded cart fixture, run attempts one and two with a controlled overlap, then compare event order and terminal state against matched success cases.

    Stop rules

    Stop immediately if any unredacted sensitive field appears, if cardinality exceeds the agreed cap, if the test affects a real customer, or when both hypotheses are falsified or one is reproduced three controlled times.

    Validator

    The plan contains all ten evidence domains, two falsifiable hypotheses, a complete instrumentation contract, protocol, and stop rules and passes the bundled structural validator. Nothing was deployed or executed.

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

    About this skill

    ### More logs do not automatically create better evidence Broad debug logging can add noise, privacy exposure, high cardinality, and cost without separating competing causes. Elusive incidents need a precise symptom, a known-good comparator, falsifiable hypotheses, and the smallest safe observation that changes the decision. Observability Reproduction Designer maps correlation, time, cohort, input, application state, dependency state, retries, errors, version, and privacy. It defines temporary instrumentation with fields, location, sampling, retention, redaction, cardinality limit, and removal condition, then writes a controlled reproduction protocol and stop rules. Built for SRE, platform, backend, and incident-response teams handling intermittent, tenant-specific, timing-sensitive, or dependency-related failures. It profiles the existing trace, log, metric, privacy, test-harness, retry, and release conventions before recommending any new signal. The package includes full Skill source, evidence-map method, worked checkout incident, deterministic validator, mutation tests, and a six-page Buyer Guide PDF. ### Boundaries It does not access production, deploy telemetry, run load, collect private customer data, or claim a root cause. A complete plan is not a reproduced incident.

    Requires buyer-supplied incident notes, logs, traces, metrics, topology, deploy history, or privacy constraints. It does not access production, deploy telemetry, run load, collect private customer data, or claim a root cause. A complete plan is not a reproduced incident.

    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

    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 2 months ago

    Permissions

    No special permissions declared or detected

    Listed2 months ago

    What's inside

    Frequently Asked Questions

    This skill automates the design of high-precision telemetry plans and reproduction protocols for intermittent production bugs, helping you move from "it works on my machine" to a falsifiable test case.

    You receive the complete Skill source code, the evidence-map methodology, a worked checkout incident example, a deterministic validator, mutation tests, and a comprehensive six-page Buyer Guide PDF.

    No, it is designed for safety; it generates the instrumentation blueprints and reproduction logic, but it does not directly access your production environment, deploy code, or collect private user data.

    The skill is designed to profile your specific trace, log, and metric conventions first, ensuring that its recommendations integrate seamlessly with your existing observability stack and release workflows.

    Installation involves importing the skill source into your preferred AI agent environment; the included Buyer Guide provides step-by-step instructions for integrating the designer into your incident response workflow.

    Buyers receive lifetime access to the version purchased, including any documentation updates or refinement of the included reproduction protocols and evidence-map templates.

    New in DevOps & Cloud