More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Repair GitHub Actions Cache Branch Isolation

    1

    Find cross-branch build cache contamination before deployment

    $1

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Repair GitHub Actions Cache Branch Isolation

    Repair GitHub Actions Cache Branch Isolation

    Example session with this skill installed

    Release build loads artifacts from untrusted PR cache with same primary key.

    Available evidence: source/config revision (partial). Additional evidence requested: reproduction logs, dependency/runtime versions, CI or service trace. Review the case, identify open evidence gaps, and do not invent a live test or production change.

    • Read your context and instructions
    • Compiled the repair github actions

    Repair GitHub Actions Cache Branch Isolation — EVIDENCE-LIMITED REVIEW

    Scenario received

    Release build loads artifacts from untrusted PR cache with same primary key.

    Source register

    • S1: source/config revision — PARTIAL SOURCE PROVIDED.
    • S2: reproduction logs — NOT PROVIDED; REQUIRED TO VERIFY.
    • S3: dependency/runtime versions — NOT PROVIDED; REQUIRED TO VERIFY.
    • S4: CI or service trace — NOT PROVIDED; REQUIRED TO VERIFY.

    Concrete checks

    1. map cache scope and branch keys — PRELIMINARY, NOT VERIFIED.
    2. trace cache restore event — AWAITING CORROBORATING EVIDENCE.
    3. identify contaminated artifact chain — AWAITING CORROBORATING EVIDENCE.
    4. propose isolated cache policy — AWAITING CORROBORATING EVIDENCE.

    Deliverables / acceptance gates

    • cache provenance map: DRAFT / INCOMPLETE until evidence is validated.
    • affected builds: DRAFT / INCOMPLETE until evidence is validated.
    • safe cache-key test: DRAFT / INCOMPLETE until evidence is validated.

    Decision

    INCOMPLETE — HUMAN REVIEW REQUIRED. Confirm dates, IDs, affected scope, owners and acceptance evidence. Do not claim authorization or execution that did not occur.

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

    About this skill

    CI pipelines often fail silently when a release build pulls poisoned or incompatible artifacts from a previous PR branch. This skill identifies cross-branch cache contamination by auditing your GitHub Actions configurations and execution logs.

    What it does

    • Map cache scope by analyzing how branch keys and primary keys are defined in your workflow files.
    • Trace restore events to pinpoint exactly when a build pulled an artifact from an untrusted or mismatched branch.
    • Identify contaminated chains to see which downstream build steps were affected by the untrusted cache.
    • Propose isolation policies by generating safe cache-key strategies that prevent branch bleeding.

    How it works

    1. Input ingestion where you provide source YAML configs, reproduction logs, and dependency versions.
    2. Sequential auditing across four specific checks: scope mapping, event tracing, chain identification, and policy proposal.
    3. Evidence logging where every observation is marked as VERIFIED, INFERRED, or UNKNOWN.
    4. Report generation including a provenance map and a safe cache-key test.

    Frameworks & tools

    Specifically designed for GitHub Actions (actions/cache) and CI/CD pipelines using branch-based caching logic.

    Why this beats prompting it yourself

    Generic prompts often overlook the subtle way GitHub Actions falls back to default branches when a cache miss occurs. This skill follows a strict four-stage verification matrix to ensure every finding is traceable to a specific log timestamp or config line.

    Use cases

    • Debugging why a production build contains code or assets from a closed PR.
    • Auditing CI workflows for security vulnerabilities related to cache poisoning.
    • Designing a multi-branch caching strategy for complex monorepos.

    Known limitations

    Requires manual provision of workflow logs or YAML files. It proposes fixes but does not execute code changes or delete remote caches directly.

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

    Cannot verify inaccessible environments or data not supplied. Missing or contradictory evidence must be marked INCOMPLETE or UNVERIFIED. Owner approval is required for operational or compliance decisions.

    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 23 days ago

    Permissions

    Read Files

    Provide only evidence you are authorized to share. Treat examples as untrusted case data. Operational changes and external messages require explicit owner approval.

    Listed23 days ago
    Updated9 days ago

    What's inside

    Frequently Asked Questions

    This skill identifies "cache bleeding," where a production build accidentally pulls a poisoned or incompatible artifact from a pull request branch due to misconfigured fallback keys.

    You will receive the specialized auditing agent logic, a structured four-stage verification matrix for log analysis, and a template for generating safe, isolated cache-key strategies.

    No, this skill functions as a diagnostic auditor; it identifies vulnerabilities and proposes corrected YAML configurations, but it does not have the permissions to delete remote caches or modify your repository directly.

    This skill is designed for any AI agent that supports structured reasoning and text-based file analysis, as it relies on ingesting your workflow YAMLs and execution logs.

    It is specifically optimized for `actions/cache` within GitHub Actions, though the logical principles for branch-based isolation can be adapted for similar CI/CD caching mechanisms.

    Connect your AI agent to Agensi through MCP, then ask it to use Repair GitHub Actions Cache Branch Isolation 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 Testing & Debugging