Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    agentreceipt

    1

    Rigorous Git-based audit tool to review AI agent code changes for secrets, scope drift, and weakened tests.

    $9

    /mo

    Subscription · cancel anytime

    Secure checkout via Stripe

    • Always the latest version
    • Delivered live to your agent via MCP
    • Cancel anytime, access ends at period end
    Security scanned
    agentreceipt

    agentreceipt

    Example session with this skill installed

    Review the changes in branch 'feature/auth-fix' against the main branch. The goal was to update the token expiration logic in the login controller.

    • Read your context and instructions
    • Compiled the agentreceipt

    I have completed the review of branch feature/auth-fix against main. The change is focused but contains one high-severity security finding and two warnings regarding test coverage.

    AgentReceipt Review Report

    SeverityCategoryFileFindingQuote/Evidence
    CRITICALSecretssrc/auth/provider.tsHardcoded mock API key found in provider config.const MOCK_KEY = "sk_t... (32 chars)"
    WARNINGTeststests/login.test.tsAssertion weakened from exact match to truthy check.- expect(res.expiry).toBe(3600); + expect(res.expiry).toBeTruthy();
    WARNINGSensitive.github/workflows/ci.ymlUnexpected change to CI timeout settings.timeout-minutes: 60 # was 20
    NOTEScopesrc/controllers/login.tsLogic correctly implements 1-hour expiration.res.cookie('token', t, { maxAge: 3600000 })

    Verdict: CHANGES BLOCKED

    The presence of a hardcoded secret in src/auth/provider.ts requires immediate remediation. Additionally, the weakening of the expiry test in tests/login.test.ts reduces confidence in the fix.

    Checked: Scope, Secrets, Tests, Dependencies, Sensitive Areas.
    Could not verify: Trustworthiness of the jwt-decode upgrade (v3.1.2 to v4.0.0).

    Next steps

    1. Rotate the leaked API key immediately; it is now in the Git history.
    2. Revert the test assertion in tests/login.test.ts to use specific values.
    3. Explain why the CI workflow timeout was tripled in ci.yml.
    $9/mo

    Subscription · always the latest version

    • Secure checkout via Stripe
    • Cancel anytime
    • Delivered live via MCP

    About this skill

    AI coding agents often introduce subtle regressions, accidentally disable linting, or leak credentials during automated refactors. AgentReceipt provides a structured audit process to review AI-generated code changes before they hit production. It treats the agent's output as untrusted and applies a rigorous checklist focused on security, stability, and scope.

    What it does

    • Scope validation identifies drift where the agent modified files or logic unrelated to the original ticket.
    • Secret detection scans for hardcoded API keys, tokens, or passwords accidentally committed to history.
    • Test integrity checks flag instances where the agent skipped, deleted, or weakened existing test assertions to make code pass.
    • Dependency monitoring highlights new packages or version overrides that could introduce supply chain risks.
    • Safety auditing triggers alerts when changes touch sensitive areas like authentication, payments, or CI configurations.

    How it works

    1. Identify the target by specifying a branch, pull request, or commit range for the audit.
    2. Define intent by providing the original prompt or ticket description to evaluate against the actual diff.
    3. Analyze facts using read-only Git commands to map file statuses, diff stats, and commit logs.
    4. Execute checklist following a strict order of operations from scope and secrets to tests and dependencies.
    5. Generate report based on standardized templates, grading every finding by severity with cited evidence.

    Frameworks & tools

    This skill is framework-agnostic and works with any project managed by Git. It relies on standard Git CLI tools like git diff, git log, and git merge-base.

    Why this beats prompting it yourself

    Generic prompts often miss "silent" failures like a linter being disabled via a comment or a test being changed from strict to loose. This skill follows a deterministic, file-backed methodology that requires cited quotes for every finding, preventing hallucinated bugs while ensuring no security checklist items are skipped.

    Use cases

    • Audit a large refactor performed by an automated agent for scope creep.
    • Scan PRs for accidentally committed environment variables or cloud provider keys.
    • Verify that an agent didn't "cheat" on a bug fix by commenting out failing unit tests.

    Known limitations

    Cannot verify the supply chain trustworthiness of new dependencies. Requires a local Git environment and cannot run code or tests unless explicitly requested by the user.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Code Quality & Review.

    Cannot verify dependency trustworthiness. Requires local Git environment. Does not execute code or tests by default. Assessment of intent requires user-provided context.

    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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 7 days ago

    Software
    Git

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Listed7 days ago

    What's inside

    Frequently Asked Questions

    AgentReceipt acts as a rigorous automated auditor that specifically looks for "agent-only" mistakes like scope drift, hidden hardcoded secrets, and the deletion or weakening of existing test assertions that manual reviewers might miss during large refactors.

    This skill is framework-agnostic and works with any AI agent capable of executing Git commands; it has been tested with popular coding agents like Aider, OpenDevin, and custom GitHub Action agents.

    No, AgentReceipt uses read-only Git commands and localized analysis to audit code changes without ever executing the code or running tests, ensuring your environment remains secure during the review process.

    Installation is a simple plug-and-play process; you simply provide your agent with access to the skill and your local Git environment, then point it at a specific branch or commit range to begin.

    Yes, the skill is designed to compare the agent's actual code changes against the original ticket or prompt, flagging "scope creep" where the agent modified files or logic unrelated to the requested task.

    Your purchase includes the full skill definition, a set of standardized audit report templates, and a deterministic checklist methodology that ensures every finding is cited with evidence from the Git diff.

    New in Code Quality & Review