Test Quarantine Recovery Ledger

    1

    Keep original CI failures visible after retries. Join attempt evidence with quarantine owners and expiry dates, then produce a review ledger with clear gaps and next steps.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Test Quarantine Recovery Ledger

    Example session with this skill installed

    Use Test Quarantine Recovery Ledger on the included synthetic examples/attempts.json and examples/quarantines.json as of 2026-11-06. Preserve original failures, identify missing evidence, and return the complete Markdown reading copy. Do not change any quarantine settings.

    • Read your context and instructions
    • Compiled the test quarantine recovery

    Local verified synthetic example, not a real CI run or an Agensi execution. The exact uploaded package produced the complete Markdown output below; no output is omitted.

    Test Quarantine Recovery Ledger

    As of 2026-11-06

    Passing on retry is not proof of flakiness. No quarantine removal is authorized.

    Sample summary

    • Executions: 7
    • First-attempt outcomes: {"cancelled": 0, "error": 1, "failed": 2, "passed": 2, "skipped": 1, "unknown": 0}
    • Missing original attempts: 1
    • Comparable first failure then last observed pass: 1
    • Incomplete or incomparable sequences: 2
    • Coverage assertion: {"complete": false, "description": "Synthetic selected runs; not a full CI export", "end": "2026-11-05", "start": "2026-11-01"}

    auth::refresh::chromium

    • Owner: Identity QA
    • State: active; expires: 2026-11-06
    • Metadata flags: expires_today
    • Executions: 2
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 0, "passed": 2, "skipped": 0, "unknown": 0}
    • Comparable first failure then last observed pass: 0
    • Next: Review metadata flags
    • Next: Review observed passes against required coverage and release criteria
    • Execution 8042:refresh: 1:passed; pass_observed
    • Execution 8043:refresh: 1:passed; pass_observed

    billing::invoice::chromium

    • Owner: Billing QA
    • State: active; expires: 2026-11-14
    • Metadata flags: none
    • Executions: 1
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 1, "passed": 0, "skipped": 0, "unknown": 0}
    • Comparable first failure then last observed pass: 0
    • Next: Repair missing or changed comparison evidence before interpreting retries
    • Next: Investigate failure/pass transitions; classify root cause separately
    • Next: Inspect failure evidence and final outcomes
    • Execution 8045:invoice: 1:failed -> 2:passed; incomplete_or_incomparable

    carousel::mobile::chromium

    • Owner: Web QA
    • State: active; expires: 2026-11-15
    • Metadata flags: none
    • Executions: 1
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 0, "passed": 0, "skipped": 1, "unknown": 0}
    • Comparable first failure then last observed pass: 0
    • Next: Obtain decisive outcomes; skipped, cancelled and unknown are not passes
    • Execution 8047:carousel: 1:skipped; no_decisive_last_observed_outcome

    checkout::visa::chromium

    • Owner: Payments QA
    • State: active; expires: 2026-11-05
    • Metadata flags: expired
    • Executions: 1
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 1, "passed": 0, "skipped": 0, "unknown": 0}
    • Comparable first failure then last observed pass: 1
    • Next: Owner review of expired quarantine; do not auto-release
    • Next: Review metadata flags
    • Next: Investigate failure/pass transitions; classify root cause separately
    • Next: Inspect failure evidence and final outcomes
    • Execution 8041:checkout: 1:failed -> 2:passed; first_failure_then_last_observed_pass

    exports::pdf::chromium

    • Owner: Reports QA
    • State: active; expires: UNSET
    • Metadata flags: missing_expires_on
    • Executions: 0
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 0, "passed": 0, "skipped": 0, "unknown": 0}
    • Comparable first failure then last observed pass: 0
    • Next: Review metadata flags
    • Next: Obtain current execution evidence; no observations is not a pass

    reports::csv::chromium

    search::autocomplete::chromium

    • Owner: UNASSIGNED
    • State: active; expires: 2026-11-13
    • Metadata flags: missing_owner
    • Executions: 1
    • First-attempt outcomes: {"cancelled": 0, "error": 0, "failed": 0, "passed": 0, "skipped": 0, "unknown": 0}
    • Comparable first failure then last observed pass: 0
    • Next: Assign an accountable owner
    • Next: Review metadata flags
    • Next: Repair missing or changed comparison evidence before interpreting retries
    • Next: Review observed passes against required coverage and release criteria
    • Execution 8044:search: 2:passed; incomplete_or_incomparable

    Limitations

    • Passing on retry is an observation, not proof of flakiness or root cause.
    • This report never authorizes removing a quarantine or changing CI gates.
    • Only supplied rows are analyzed; omitted runs, suppressed tests, final-only exports, and selection bias cannot be detected reliably.
    • Test keys must encode suite, parameterization, shard/variant and environment as needed. Execution IDs must identify one invocation and its retries, not a whole workflow.
    • Attempt 1 is the original test execution; workflow reruns are not automatically test retries. Exporters must preserve this distinction.
    • Revision and environment equality are necessary comparisons, not proof that all relevant conditions were identical.
    • Failure signatures and evidence URLs are caller-supplied, unverified data. URLs are not fetched. Report contents may be private.
    • As-of is used only for quarantine dates; it does not filter attempts or reconstruct historical metadata.
    • Coverage is a caller assertion. Counts describe this sample, not a population failure probability; no confidence claim is made.

    Input receipt

    • {"bytes": 5206, "name": "attempts.json", "role": "ci_attempts", "sha256": "fa11a6aa30efb492e3343c3bc99d8a47d93aed6eda7e18ad8abd3e27c64549fe"}
    • {"bytes": 2206, "name": "quarantines.json", "role": "quarantine", "sha256": "81ebfb142927efe4bc3f618b3139fdd10f6793e50696d22d229a5883ddcc18f8"}

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

    About this skill

    What it does

    A green retry can hide the failure that put a test in quarantine. This skill preserves each supplied attempt and makes the review evidence easier to use.

    Give it a normalized CI attempt export and current quarantine metadata. It separates original failures from later passes, flags missing or changed comparison context, and lists expired, ownerless and unobserved quarantines. JSON keeps the complete evidence record. Markdown provides a reading copy with review actions.

    The included synthetic example covers a passing retry, missing original attempt, code change between attempts, intermediate pass followed by failure, a skipped test and a quarantine with no observations.

    Passing on retry does not prove flakiness. The skill does not diagnose root cause, change CI gates, remove quarantines or declare tests safe to release.

    Source and license

    Author: Eugene (Agensi creator Evgeny Tkachenko). Copyright (c) 2026 Eugene. MIT license included for the newly created source, documentation and synthetic fixtures. Version 1.0.0. This is a newly created local package; no public GitHub repository or commit is claimed.

    Requirements

    Requirements: Python 3.10+, permission to read supplied files, run a temporary standard-library script, and write requested outputs. No account connection, API key, third-party dependency, install step or network access.

    Verification

    Verification: the exact uploaded ZIP passed 47 regression tests and reproduced the full synthetic example outputs on local Python 3.12.15. No real CI system or production dataset was used. On November 6, 2026, creator-account Agensi MCP delivery was verified. The delivered scripts matched the release ZIP and ran unchanged locally on macOS with Python 3.12.15. This was not Agensi-hosted execution, a non-creator buyer test or production validation. The delivery test checked JSON and Markdown outputs, incomplete retry evidence and rejection of an invalid assessment date.

    Normalized JSON schema_version 1 or the documented CSV fields only; not a raw JUnit/XML parser or native CI connector. Missing or changed comparison context prevents comparable-retry counts. Only supplied rows are analyzed; no completeness, causality, population reliability or safe-release claim. Passing on retry does not prove flakiness. As-of evaluates current quarantine metadata, not a historical attempt cutoff. No real CI or production-data test. Verification covered creator-account MCP delivery and local execution only, not Agensi-hosted execution, non-creator access, production use or broader compatibility.

    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

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 10 days ago

    • Free with an Agensi account

    Requires

    Python 3.10+
    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files

    Reads the supplied normalized attempt and quarantine files, writes one temporary Python script and optionally new report files, and runs that standard-library script locally. No network, secrets, new dependencies, CI edits or automatic quarantine removal. Keep outputs private unless separately authorized.

    Listed10 days ago

    What's inside

    Frequently Asked Questions

    A JSON evidence ledger and a Markdown reading copy. It retains original failures and retry chains, source hashes and row references, and flags expired, ownerless or unobserved quarantines for human review.

    Normalized JSON or CSV attempt data, JSON quarantine metadata, and an assessment date. Preserve all available attempts and stable test/execution identities. Python 3.10+ and local script/file access are required. No CI account, network connection, paid tool or API key is needed.

    No. The tool organizes supplied evidence only. It neither diagnoses flakiness nor authorizes release or CI changes. Verification used synthetic fixtures. Creator-account MCP delivery and unchanged local execution were verified on November 6, 2026; no real CI system, Agensi-hosted runtime or independent buyer was tested. The free package is MIT-licensed, copyright 2026 Eugene.

    New in Testing & Debugging