Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    AET Codebase to AI Instructions

    1

    Discover coding conventions in your repository, review the evidence, and turn approved standards into reusable AI instructions.

    Free

    Security scanned
    AET Codebase to AI Instructions

    AET Codebase to AI Instructions

    Example session with this skill installed

    Extract coding standards from the /src/components directory of this React project. I want to see how we handle functional components and PropType definitions. Inline output only.

    • Read your context and instructions
    • Compiled the aet codebase to
    • Generated the document

    I have completed the discovery phase for the /src/components directory. I scanned 18 files and identified 4 candidate conventions based on recurring patterns in the functional components.

    Candidate Coding Standards Proposal [NOT PERSISTED]

    IDClassCandidate RuleRationaleProvenance
    STND-001Repeated ConventionUse arrow functions for functional component definitions.14/18 components use const X = () => syntax./src/components/Button.tsx:4, /src/components/Card.tsx:12
    STND-002Explicit PolicyAll props must be documented using TypeScript interfaces, not PropTypes.Found interface Props in 100% of new components; README.md section 2 confirms deprecation of PropTypes./src/components/List.tsx:8, /README.md:45
    STND-003Isolated ExampleUse useMemo for all object literals passed to context providers.Only observed in AuthProvider.tsx; insufficient evidence for repo-wide mandate./src/components/Auth/AuthProvider.tsx:22
    STND-004Conflicting EvidenceNaming of utility components (PascalCase vs camelCase).5 components use PascalCase, 3 use camelCase; needs manual resolution./src/components/utils/

    Recommendation Summary

    • Adopt: STND-001, STND-002.
    • Return-for-evidence: STND-004 (Requires scan of /src/hooks to see if camelCase is reserved for hooks only).
    • Abandon: STND-003 (Too specific to context providers).

    Next steps

    1. Approve IDs for adoption (Gate A).
    2. Specify the destination file for the generated instructions (Gate B).
    3. Request a batch increase if you want to resolve the STND-004 naming conflict across more directories.

    Button.tsx

    PDF · document

    Generated

    Example file from a real run - the skill writes it into your workspace.

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

    About this skill

    Manual extraction of coding standards from a large codebase is slow and prone to bias. This skill solves that by providing a systematic framework to discover, classify, and review conventions based on repository evidence rather than model preference. It ensures that every proposed standard is backed by specific file citations and explicit user approval before any instructions are generated.

    What it does

    • Evidence-backed extraction scans repositories to identify explicit policies, repeated conventions, and isolated examples.
    • Strict classification categorizes observations into five evidence classes to separate documented policy from mere recurring behavior.
    • Budget-aware discovery manages token usage by sampling representative clusters and paged inventory rather than bulk ingestion.
    • Two-gate verification requires independent user approval for both the content of a rule and its intended storage destination.
    • Conflict resolution identifies and reports contradictions between existing project instructions and observed code patterns.

    How it works

    1. Establish the Contract to define the read scope, repository identity, and intended audience for the standards.
    2. Discover and Extract candidates using defined budgets for lines and files, recording stable IDs and source provenance for every observation.
    3. Review and Gate by presenting a proposal where the user must explicitly approve specific IDs and output paths.
    4. Generate Instructions only for accepted items, creating new Markdown files or performing conflict-aware merges into existing owners.

    Frameworks & tools

    This skill is language-agnostic and works with any repository accessible via standard file read and search capabilities. It uses a structured workflow involving specific templates for proposals and approved instructions without requiring external APIs or environment telemetry.

    Why this beats prompting it yourself

    Generic prompts often hallucinate "best practices" that don't exist in your specific codebase. This skill enforces a strict reading contract that treats source code as data, preventing the AI from imposing its own stylistic biases or ignoring existing local policies.

    Use cases

    • Onboarding new developers by generating a clear guide of "how we write code here" from existing samples.
    • Standardizing legacy code by identifying fragmented conventions and proposing a single unified path.
    • Instruction engineering to create scoped system prompts for other agents based on actual repository truth.
    • Audit and compliance to verify if documented policies match the actual implementation across a project.

    Known limitations

    • Does not perform refactoring, execute project code, or mine architecture.
    • Requires authorized read access to the target repository.
    • Findings may reflect sampled evidence, not complete repository coverage.
    • Generating instruction files requires explicit adoption and separate authorization of the destination and write mode.
    • When safe writing is unavailable, outputs remain inline and unpersisted.

    Privacy and confidentiality

    This skill provides AI workflow guidance, not a confidentiality or NDA-compliance guarantee. Use it only with information you are authorized to process through your chosen AI provider. Local files may be processed by that provider; local access does not imply local-only processing. Review generated outputs before sharing or publishing.

    The skill is provided "as is," without warranties, to the extent permitted by applicable law.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.

    No code execution or refactoring. Requires explicit user gates for all writes. Budget-limited to 1,500 lines by default. Extraction-only if disclosure scope is ambiguous.

    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 20 days ago

    • Free with an Agensi account

    Permissions

    Read Files
    Write Files
    Listed20 days ago

    What's inside

    Frequently Asked Questions

    Unlike generic LLM prompts that hallucinate "best practices," this skill uses an evidence-backed extraction process that maps rules to specific file citations, ensuring all standards reflect your actual codebase rather than model bias.

    The skill is language-agnostic and works with any codebase, though its metadata and tagging are optimized for modern environments like TypeScript and React. It requires read access to your repository and a platform capable of handling structured Markdown output.

    No, this skill is designed for discovery and documentation; it will identify patterns and generate Markdown instructions, but it never modifies your source code or performs automated refactoring.

    Every purchase includes the complete workflow for the 'aet coding standards extraction' skill, including the systematic discovery framework, classification templates, and the conflict-resolution logic.

    The skill utilizes a budget-aware discovery process that uses paged inventory and representative sampling to manage token consumption, preventing errors or high costs when scanning large repositories.

    New in Prompt & Skill Engineering

    Free