Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    blog post pipeline

    1

    A gated 8-stage editorial pipeline that enforces fact-checking, brand voice, and multi-role adversarial review.

    Free

    1 useSecurity scanned
    blog post pipeline

    blog post pipeline

    Example session with this skill installed

    I need to start a new blog post about the benefits of memory-safe languages for legacy systems. The target is our 'Engineering Insights' blog. Start the brief and research stages.

    • Read your context and instructions
    • Compiled the blog post pipeline

    I have initialized the post directory and completed the research phase. The process is currently halted at the Curation gate for your approval of the gathered facts.

    Post Pipeline: Memory-Safe Languages

    StageStatusOutput FileAction
    0. BriefCompleteposts/memory-safe-legacy/00-config.mdDefined target: Engineering Insights
    1. ResearchCompleteposts/memory-safe-legacy/01-research.mdFound 12 sources, 4 primary benchmarks
    2. CurationPendingposts/memory-safe-legacy/02-curated.mdAwaiting human selection

    Research Summary

    I have categorized the findings into three key pillars. Please review the 01-research.md file and select which points should move to the drafting stage.

    Key Pillar 1: CVE Reduction Statistics

    • Source [1]: CISA report on memory safety vulnerabilities (approx 70% of deep vulnerabilities).
    • Source [2]: Google/Microsoft case studies on Rust adoption.

    Key Pillar 2: Performance Overheads

    • Source [3]: Benchmarks comparing C++ manual management vs. Rust/Swift runtime checks.
    • Source [4]: Impact analysis on legacy hardware constraints.

    Key Pillar 3: Integration Patterns

    • Source [5]: Best practices for FFI (Foreign Function Interface) with legacy C codebases.
    • Source [6]: Incremental migration strategies (strangler fig pattern).

    Next steps

    1. Open posts/memory-safe-legacy/01-research.md to review all 12 citations.
    2. Create posts/memory-safe-legacy/02-curated.md containing only the facts you want the writer to use.
    3. Update the status in 00-config.md to curated to enable the Drafting stage.

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

    About this skill

    Manual blog production often suffers from hallucinated facts, drifting brand voice, and fragmented workflows. This skill implements a rigorous, gated editorial pipeline that enforces truth and consistency across multiple publications.

    What it does

    • Enforces veracity by blocking any quantitative or comparative claims that lack a cited source from the research phase.
    • Isolates roles to prevent the writer from researching or the reviewer from self-editing, ensuring objective adversarial oversight.
    • Manages state via a file-based system, allowing you to pause and resume complex content projects across different sessions.
    • Handles localization through cultural adaptation rather than literal translation, keeping the message intact across languages.
    • Automates publishing including image planning, asset uploads, and final Markdown deployment to target repositories.

    How it works

    1. Initialize the brief by creating a configuration file that defines the blog target and primary goals.
    2. Execute research and curation where the agent gathers sources and the human selects which facts are approved for the draft.
    3. Generate and review the draft, triggering a specialized adversarial pass that flags unsourced claims as blockers.
    4. Finalize and publish by optionally translating the content, planning media, and pushing to the target environment.

    Frameworks & tools

    The skill is designed for environments with filesystem access like Claude Code or Cursor. It uses YAML for brand configuration and maintains state through a structured Markdown file hierarchy.

    Why this beats prompting it yourself

    Generic prompts often lead to "hallucinated expertise" where the AI invents statistics to sound persuasive. this skill forces a structural separation between fact-finding and prose generation, making it impossible for the agent to include unverified data.

    Use cases

    • Technical content marketing requiring precise performance benchmarks and cited specifications.
    • Bilingual blog management where posts must be adapted for different cultural contexts while maintaining a single source of truth.
    • High-volume editorial teams needing a repeatable, audited process for multi-author or multi-brand environments.

    Known limitations

    Requires initial manual setup of a blogs.yaml file for brand voice and hard limits. Does not support automated chaining of stages; a human must approve each gate to progress the state.

    Requires manual creation of blogs.yaml for brand definitions. Stages cannot be auto-chained; human approval is required at every gate. filesystem access is necessary for state management.

    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

    Requires

    Agensi
    Git

    Permissions

    Terminal / Shell

    Allowed Hosts

    www.agensi.io
    example.org

    File Scopes

    blog-post-pipeline-skill-1.0.0/**
    Listed1 month ago
    Updated1 month ago

    What's inside

    Frequently Asked Questions

    Traditional prompts often allow the AI to 'hallucinate' facts to fit a narrative; this skill uses a gated 8-stage architecture that physically separates research from writing, making it impossible for the agent to include a claim that wasn't pre-approved in the research phase.

    This skill is optimized for AI environments with filesystem access, such as Claude Code, Cursor, or local VS Code setups, as it relies on a structured Markdown and YAML file hierarchy to maintain state across sessions.

    No, the pipeline is designed for high-stakes editorial oversight, meaning a human must review and 'unlock' each gate before the agent proceeds to the next stage, ensuring total control over the final output.

    You will receive the complete editorial framework, a set of configuration templates (including the blogs.yaml schema), and a comprehensive guide on how to initialize your first project brief.

    Yes, the pipeline includes a specialized cultural adaptation stage that goes beyond literal translation to ensure your brand voice and technical nuances remain intact across different regional audiences.

    New in Agents & Orchestration