Evidence Input Manifest

    1

    Generate deterministic CSV/JSON manifests with SHA-256 hashes and metadata for local evidence files.

    Free

    1 useSecurity scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Evidence Input Manifest

    Example session with this skill installed

    Generate a CSV manifest for the ./research_papers directory. Exclude all .tmp files and include PDF page counts. Put the result in ./audit/manifest.csv.

    • Read your context and instructions
    • Compiled the evidence input manifest

    I have scanned the ./research_papers directory and generated a deterministic manifest. A total of 4 files were processed; 1 temporary file was excluded based on your glob pattern.

    Evidence Manifest (CSV)

    relative_path,size_bytes,sha256,mime_type,extension,pdf_pages,notes
    analysis_v1.pdf,124556,e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855,application/pdf,.pdf,12,
    methodology.docx,45882,5feceb66ffc86f38d952786c6d696c79c2dbc239dd4e91b46729d73a27fb57e9,application/vnd.openxmlformats-officedocument.wordprocessingml.document,.docx,,
    raw_data_summary.csv,892,b94d27b9934d3d08a52e52d7da7dabfac484efe37a5380ee9088f7ace2efcde9,text/csv,.csv,,
    supporting_images/figure1.png,2048550,ac107dc39f9b5a83a4c000c06a3c9e64e52b2cf5b8e96716a4220b2210086c9f,image/png,.png,,
    

    Next steps

    1. Verify that the SHA-256 hashes match your local source records to ensure integrity.
    2. Use the pdf_pages column to estimate the processing time for the subsequent extraction phase.
    3. Store this manifest in your version control system or audit folder to maintain a reproducibility trail.

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

    About this skill

    Manual file inventory for large evidence sets is slow and prone to error, making audit trails difficult to verify. This skill generates a deterministic manifest of local files with relative paths, sizes, and SHA-256 hashes to ensure data integrity.

    What it does

    • Deterministic indexing creates a sorted inventory of files to ensure repeatable results across different runs.
    • Integrity hashing calculates SHA-256 values for every file to verify that inputs haven't been tampered with or corrupted.
    • Metadata collection records file sizes, MIME types, extensions, and optional PDF page counts.
    • Error reporting logs unreadable files and metadata warnings without halting the entire inventory process.

    How it works

    1. Target selection where the user identifies the input directory containing the evidence and the desired output location.
    2. Execution of the manifest script using either CSV or JSON format, with options to exclude specific glob patterns.
    3. Verification of the file count and notes regarding any skipped or unreadable files.
    4. Handoff of the generated manifest to be stored alongside the raw evidence for future audit or extraction steps.

    Frameworks & tools

    This skill utilizes a Python-based utility script and requires pdfinfo if PDF page counting is requested. It is designed to run in local environments or controlled containers where file system access is available.

    Why this beats prompting it yourself

    Writing reliable hashing scripts that handle relative paths and MIME types correctly takes time to debug. This skill enforces strict reproducibility rules, ensuring environment-specific data like absolute paths or hostnames don't pollute the manifest and break deterministic checks.

    Use cases

    • Prepare a verifiable record of evidence before running extraction pipelines.
    • Audit a folder of research papers for duplicates using SHA-256 hashes.
    • Create a handoff manifest for third-party reviewers to confirm data integrity.
    • Inventory large datasets to track file types and sizes before cloud migration.

    Known limitations

    The skill inventories files only and does not extract text, interpret content, or follow URLs. PDF page counting requires the pdfinfo system utility to be installed.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Requires Python and optionally pdfinfo. Does not perform OCR or content extraction. Manifests are limited to local file system metadata.

    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

    1 use

    Times used to date

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 27 days ago

    • Free with an Agensi account

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed27 days ago

    What's inside

    Frequently Asked Questions

    This skill generates a standardized, tamper-proof inventory of your local files, providing SHA-256 hashes and metadata that ensure your data hasn't been altered between the collection and analysis phases.

    The skill is designed for local environments or secure containers with filesystem access and is compatible with any AI agent or platform capable of executing Python-based utility scripts.

    Standard metadata includes relative paths, file sizes, and MIME types; if you have the 'pdfinfo' utility installed, the skill can also automatically record page counts for PDF documents.

    No, the skill enforces strict reproducibility by using relative paths and excluding host-specific data, ensuring the manifest remains identical regardless of which machine or directory it is run from.

    You receive the production-ready Python utility, detailed integration documentation for your agent, and instructions for configuring optional system dependencies like PDF page counting.

    The skill includes a robust error-handling mechanism that logs unreadable files or permission issues in a summary report without crashing the entire indexing process, allowing you to address specific files later.

    New in LLM Ops

    Free