Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Instinct

    2

    Audit repository history to find undocumented breaking changes and missing engineering trade-offs.

    $12.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Instinct

    Instinct

    Example session with this skill installed

    Audit the last 15 commits for engineering provenance and breaking changes. Focus on the public API and shared config files.

    • Read your context and instructions
    • Compiled the instinct

    I have audited the last 15 commits in the repository history, filtering for changes to the public API, shared schemas, and trust boundaries. Four commits met the criteria for Critical-tier-shaped changes.

    Engineering Provenance Audit

    Commit HashSummaryProvenanceVerification
    a7d2f91Remove deprecated /v1/auth endpointMissingPresent
    f8e32c4Update validation schema for user-serviceMissingMissing
    b1a9d82Narrow default timeout in fetch-clientPresentMissing
    e45c710Rename Export Interface for PluginLoaderPresentPresent

    Summary: 4 qualifying commits identified in the last 15 changes.

    Finding Details

    f8e32c4: This commit modified the shared JSON schema used by three downstream services. The commit message only states "update schema" without explaining why the minItems constraint was increased. No unit tests were added to the schema validator in this diff.

    a7d2f91: While the breaking change was expected, there is no record of why the deprecation period was shortened instead of waiting for the Q4 milestone. The verification is present via the removal of integration tests, but the "why" is lost to history.

    b1a9d82: The timeout change has a high blast radius for consumers on slow networks. The commit message mentions the tradeoff for resource exhaustion but lacks a corresponding test case for the new limit.

    Next steps

    1. Add a regression test for the fetch-client timeout in the next PR.
    2. Document the schema validation constraints in the internal wiki or as a comment in user-service.
    3. Verify if any active clients are still hitting the now-removed /v1/auth endpoint.

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

    About this skill

    Maintaining engineering standards across a fast-moving repository is difficult when team members bypass documentation for speed. This skill identifies where critical reasoning was lost in your project history. It audits repositories to find undocumented breaking changes and missing architectural trade-offs that lead to technical debt.

    What it does

    • Identifies critical changes by scanning git history for renamed public functions, modified endpoints, or narrowed configuration defaults.
    • Evaluates commit provenance by checking if commit messages explain specific trade-offs rather than just describing the diff.
    • Surfaces missing verification by flagging high-impact changes that lack corresponding updates to tests or assertions.
    • Audits policy shifts by detecting changes to shared configurations, schemas, or instructions that affect multiple consumers.
    • Prioritizes findings by ranking commits missing both reasoning and verification at the top of the report.

    How it works

    1. Scan the recent git history or a specific range provided by the user.
    2. Filter for commits that meet the "Critical-tier-shaped" criteria, such as public API changes or trust-boundary modifications.
    3. Analyze each qualifying commit for evidence of alternatives weighed and verification artifacts added.
    4. Generate a one-shot report summarizing the provenance and verification status for each flagged commit.

    Frameworks & tools

    This skill operates directly on git history and diffs. It is compatible with any repository structure using standard git version control, including monorepos and distributed microservices.

    Why this beats prompting it yourself

    Generic prompts often miss the "blast radius" of config-only changes or fail to distinguish between descriptive and trade-off-based commit messages. This skill uses a specific rubric to distinguish between routine code churn and high-risk architectural shifts that require documented reasoning.

    Use cases

    • Perform a technical debt audit before a major version release.
    • Verify that a new team member is following the project's documentation standards.
    • Identify undocumented breaking changes in a shared library or internal SDK.
    • Conduct a post-mortem on a period of high-velocity development to find gaps in reasoning.

    Known limitations

    This skill only reads git history and diffs. It cannot rewrite commit messages or automatically patch missing tests. Finding live bugs requires a separate explicit request.

    Reads git history and diffs only. Cannot rewrite history or propose code fixes. Scope is limited to identifying gaps in documentation and verification, not fixing the underlying bugs.

    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 1 month ago

    Requires

    Groq
    Python
    Instinct MCP

    Permissions

    Terminal / Shell
    Read Files
    Environment Variables

    Allowed Hosts

    api.groq.com
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    Instinct Audit scans your git history to identify "lost" engineering context, specifically flagging high-impact changes that lack documented trade-offs or corresponding test updates. It helps teams prevent technical debt by surfacing undocumented breaking changes that typical code reviews might miss during high-velocity sprints.

    This skill is compatible with any AI agent or development environment that has access to a standard Git CLI. It works across all programming languages and repository structures, including monorepos and microservices, as it analyzes commit patterns rather than specific syntax.

    No complex integration is required; simply provide the skill with access to your repository's git history or a specific commit range. It operates directly on diffs and commit metadata to generate a report without needing to execute your code or access external databases.

    Instinct Audit is a "read-only" diagnostic tool that identifies gaps in your project's reasoning and verification history. While it surfaces where documentation or tests are missing, it will not modify your code, rewrite commit messages, or automatically generate patches.

    You receive the full skill package including the audit rubric and logic required to run the analysis on your agent. This includes lifelong access to the specific version purchased, ensuring you can run audits across multiple projects and teams.

    New in Code Quality & Review