More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Sdd Kit

    10

    Independent consistency analyst that audits specs, ADRs, and plans for contradictions and risks before implementation.

    Free

    21 uses5.0 (10 reviews)
    sdd-kit

    sdd-kit

    Example session with this skill installed

    Analyze the new authentication spec in docs/specs/auth-v2.md against our ADRs and the current sprint tasks in tasks.md. Check for contradictions and testability.

    • Read your context and instructions
    • Compiled the sdd-kit

    I have completed the consistency analysis for the Authentication V2 specification. I cross-referenced auth-v2.md with the Architecture Decision Records (ADR-004 to ADR-009) and the current task backlog.

    Consistency Analysis Report

    IDSeverityCategoryFindingImpact
    B-01BlockerContradictionSpec auth-v2.md proposes JWT-only sessions, but ADR-007 mandates secure cookie-based sessions for web clients.Security policy violation.
    M-01MajorTestabilityAC #4 "System must be fast" is subjective and lacks a millisecond latency threshold.Criteria cannot be validated in CI.
    M-02MajorOmissionNo rollback procedure defined for the database schema migration listed in tasks.md.High risk during deployment failure.
    m-01MinorCoverageTask #104 mentions Redis, but Redis implementation is not listed in the technical dependencies of the spec.Minor documentation drift.
    m-02MinorStale RefSpec refers to 'UserStore' which was renamed to 'AccountRepository' in the last refactor.Potential developer confusion.

    Next steps

    1. Update auth-v2.md to align with ADR-007 regarding cookie-based session management.
    2. Define a concrete latency threshold (e.g., < 200ms) for the authentication flow to satisfy AC #4.
    3. Add a rollback task to tasks.md for the SQL migration.
    Authentication V2 specification. I cross-referenced auth-v2.md

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

    About this skill

    Specifications and planning documents often drift from the actual codebase, leading to implementation errors and architectural debt. This skill acts as an independent consistency analyst to identify contradictions, omissions, and untestable criteria before you write a single line of code. ## What it does - **Conflict Detection** identifies contradictions between the project constitution, technical specs, and recent architectural decisions. - **Coverage Auditing** verifies that all acceptance criteria and decision records are reflected in the current implementation tasks. - **Risk Assessment** flags missing rollback plans, security oversights, and operational check omissions. - **Testability Review** ensures that requirements are specific enough to be verified through automated or manual testing. - **Blocker Identification** categorizes findings by severity to prevent high-risk implementation starts. ## How it works 1. Provide the current specification, decision logs, and task list for the upcoming sprint or feature. 2. The skill scans the repository evidence to check for stale revisions or alignment issues. 3. It generates an analysis report classifying issues as Blocker, Major, or Minor. 4. Review the report to resolve high-priority findings before proceeding to the implementation phase. ## Frameworks & tools This skill is framework-agnostic and works with any text-based documentation systems including Markdown specs, Architecture Decision Records (ADR), and structured task lists in repositories. ## Why this beats prompting it yourself General purpose prompts often miss the nuance of stale revisions or fail to cross-reference multiple sources like the project constitution and specific decision logs. This skill enforces a systematic audit across all project artifacts to ensure no contradiction goes unnoticed. ## Use cases - Audit a new feature spec against the existing project constitution for alignment. - Review a task list to ensure all acceptance criteria are covered by specific items. - Check if a complex architectural change has a defined rollback and security plan. - Verify that requirements in a PR description are actually testable. ## Known limitations Does not modify application code or automatically repair artifacts. Requires documentation to be in a readable text format within the repository.

    Cannot modify code or fix documents. Requires existing specs and decision logs to function. Analysis is limited to provided text-based repository evidence.

    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

    5.0
    10 reviews
    5
    10
    4
    0
    3
    0
    2
    0
    1
    0

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Git

    Permissions

    Terminal / Shell
    Listed1 month ago

    What's inside

    Frequently Asked Questions

    The skill requires read-access to your project's text-based documentation, such as Markdown specifications, ADRs, and task lists. It does not need access to your source code, as its primary function is to audit the logic and consistency of your planning documents.

    Unlike general prompts, this skill uses a specialized multi-source cross-referencing logic designed to catch 'drift' between old constitutions and new specs. It specifically looks for contradictions and untestable criteria that generic AI interactions typically overlook.

    The skill outputs a structured analysis report that categorizes findings into Blocker, Major, and Minor risks. This allows your team to immediately identify which architectural contradictions must be fixed before the sprint begins.

    Yes, 'sdd analyze' is framework-agnostic and works with any documentation style as long as it is in a readable text format. It is particularly effective for teams using ADRs (Architecture Decision Records) and structured Acceptance Criteria.

    This skill is designed to integrate with any AI agent that has repository read permissions. Once added to your agent's toolkit, you can trigger an audit by simply pointing the agent to the relevant documentation folders.

    New in Code Quality & Review