Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Explain My PR (ADHD)

    1

    Generates plain-language, manager-friendly documentation for Git merge requests based on code diffs.

    Free

    1 useSecurity scanned
    Explain My PR (ADHD)

    Explain My PR (ADHD)

    Example session with this skill installed

    Explain the changes in my feat/auth-fix branch where I added a retry loop to the token middleware. Explain it for my lead who hasn't seen the code yet.

    • Read your context and instructions
    • Compiled the explain my pr

    I have analyzed the diff between main and feat/auth-fix. I generated the explanation document at docs/mr-change-explanations/auth-middleware-retry.md.

    Why auth-middleware-retry changes when merging feat/auth-fix into main

    This MR modifies the token validation logic in the middleware layer. It affects 1 file with 14 additions and 2 deletions.


    What is this and why does it matter?

    • Token Middleware: Intercepts all incoming API requests to validate JWTs.
    • Resilience: Prevents 401 errors caused by transient network blips to the identity provider.

    TL;DR

    #What changedWhy it matters
    1Added exponential backoffEliminates request failure on 503 IdP responses.

    Big picture

    [Client] -> [Middleware] -> [Identity Provider]
    
                      |                ^
    
                      +-- (Retry 3x) --+
    

    Fix 1 — Exponential backoff for token validation

    What problem this fixes

    When the identity provider is under heavy load, it returns a 503. The current code immediately fails with a 401 Unauthorized, forcing users to refresh their browser.

    The change in code

    main (before)

    src/auth/middleware.py:12-15 (main branch):

    def validate(token):
        # Directly calls IdP without retry
        return idp_client.verify(token)
    
    feat/auth-fix (after)

    src/auth/middleware.py:12-22 (feat/auth-fix branch):

    @retry(stop=stop_after_attempt(3), wait=wait_exponential())
    def validate(token):
        # Added retry decorator to handle transient IdP errors
        return idp_client.verify(token)
    

    Why this is a safe change

    The decorator only catches IdentityServerUnavailable exceptions. All other auth failures (invalid tokens, expired keys) still fail immediately.


    Impact table

    ConsumerImpact
    API ClientsReduced 401 errors during peak load.
    IdP ServiceMinor increase in request volume during outages.

    Reviewer checklist

    • Grep wait_exponential to ensure max delay is capped at 2 seconds.
    • Verify idp_client.verify is idempotent.

    Suggested commit message

    feat(auth): add exponential backoff to token validation
    
    Implements a 3-attempt retry loop for token verification to handle
    transient 503 responses from the identity provider.
    

    Next steps

    1. Review the generated Markdown file in your docs/ folder.
    2. Copy the "Suggested commit message" into your git commit.
    3. Attach this document to your Merge Request description.

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

    About this skill

    Reviewers often lack the time to parse complex diffs, leading to rubber-stamp approvals or long merge delays. This skill generates manager-friendly documentation that explains the why behind your code changes. It transforms technical git diffs into structured, plain-language reports that highlight impact, safety, and business logic rather than just syntax.

    What it does

    • Drafts MR summaries by analyzing git diffs and logs to produce high-level narrative explanations.
    • Generates before/after comparisons with branch-specific labels, line ranges, and annotated code snippets.
    • Visualizes call flows using ASCII architecture diagrams to show how logic changes across the system.
    • Creates impact tables that explicitly define what remains unchanged for existing consumers versus new opt-in behavior.
    • Constructs reviewer checklists with specific grep commands and verification steps to speed up the approval process.

    How it works

    1. Trigger the skill by providing a git diff or asking to "explain this MR" for a specific branch or feature.
    2. The agent gathers git facts using git diff --stat and git log to identify the scope of changes.
    3. It produces a structured Markdown document saved to docs/mr-change-explanations/ following a standardized management template.

    Frameworks & tools

    Works with any language or framework accessible via standard Git CLI. Outputs are standard GitHub-flavored Markdown.

    Why this beats prompting it yourself

    Generic prompts often result in a list of file changes without context. This skill enforces a rigorous structure including safety assessments, ASCII flow diagrams, and side-by-side code citations that map directly to your branch names and line numbers.

    Use cases

    • Explaining complex refactors to non-technical product managers.
    • Documenting critical bug fixes for compliance or audit trails.
    • Onboarding new team members to a specific feature's logic.
    • Standardizing merge request quality across a distributed engineering team.

    Known limitations

    Requires access to git history and diffs to provide accurate line ranges. Best suited for single logical units or cohesive features rather than massive multi-repo migrations.

    Requires local git access for accurate diffs and logs. Designed for cohesive feature branches; extremely large, unrelated diffs should be split into multiple documents.

    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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    This skill transforms technical Git diffs into structured, plain-language reports, helping non-technical managers understand the business impact of code changes while providing reviewers with clear ASCII call flows and verification checklists.

    The skill works with any programming language or framework, as long as the project uses Git for version control and the agent has access to the standard Git CLI.

    Upon purchase, you receive the full skill package including the automated documentation template, the ASCII architecture visualization engine, and a set of pre-configured reviewer checklist generators.

    This skill is designed to run within any AI agent environment that supports filesystem access and Git CLI integration, outputting standardized GitHub-flavored Markdown files.

    Setup is minimal: once the skill is added to your agent, you simply provide a git diff or branch reference, and it automatically handles the analysis and document generation in your specified directory.

    While the skill is highly effective for feature work and refactors, it is currently optimized for single-repository logical units rather than massive multi-repo migrations or monorepo-wide architectural shifts.

    New in Code Quality & Review