Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Dispatch Delivery Evidence Judge — Verify Message Receipt

    1

    Capture terminal history and watched-file hashes before and after a message. The judge separates screen and file evidence, excludes unsent prompt text, and never treats an already-busy baseline as confirmed delivery.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Dispatch Delivery Evidence Judge — Verify Message Receipt

    Dispatch Delivery Evidence Judge — Verify Message Receipt

    Example session with this skill installed

    I sent "Prepare the weekly summary now" to my other agent. Compare the before and after snapshots and tell me whether it was really delivered.

    • Read your context and instructions
    • Compiled the dispatch delivery evidence

    I ran the judge on the bundled before/after fixture (read-only):

    $ python3 scripts/delivery_evidence_judge.py judge --before examples/before.json --after examples/after.json --message "Prepare the weekly summary now" --json
    {
      "exit_code": 0,
      "files": {
        "added": [],
        "changed": [
          "examples/delivery_record.json"
        ],
        "evidence": true,
        "reason": "content_digest_changed",
        "removed": [],
        "unchanged": []
      },
      "screen": {
        "after_matches": 1,
        "baseline_matches": 0,
        "evidence": true,
        "reason": "message_history_line_increased",
        "state": "SCREEN_EVIDENCE"
      },
      "success_requires": "new screen evidence and changed watched-file evidence",
      "verdict": "DELIVERED"
    }
    

    Summary

    Verdict: DELIVERED (exit code 0).

    EvidenceResultReason
    ScreenyesThe message appears as a new history line (0 matches before, 1 after)
    Watched fileyesexamples/delivery_record.json changed by SHA-256

    Both independent sources agree, which is what DELIVERED requires.

    How to read the other verdicts

    VerdictExitMeaning
    NO_EVIDENCE1Neither source changed
    SCREEN_EVIDENCE_ONLY2The message appeared, but the watched files did not change
    FILE_EVIDENCE_ONLY3Watched files changed, but the pane shows no new message
    INCONCLUSIVE_BASELINE_BUSY4The baseline was already busy, so activity cannot be attributed to this send

    Next steps

    • Add one --watch per file the receiving agent should change.
    • Capture the baseline immediately before the send and the second snapshot right after.
    • A changed hash is evidence of a content change, not proof the message caused it.

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

    About this skill

    A busy terminal is not proof that a message arrived. This read-only utility captures a pane and hashes selected files at two points, then compares transcript history and file contents independently.

    The snapshot command can capture a terminal multiplexer pane or read a saved text capture for offline use. Repeat --watch for each relevant file. The judge command checks whether the sent message appears as a new history line, excludes text left in the final prompt, detects a busy baseline as inconclusive, and compares watched files by SHA-256 rather than modification time alone.

    Verdicts

    It reports DELIVERED only when new screen evidence and a watched-file change are both present. SCREEN_EVIDENCE_ONLY, FILE_EVIDENCE_ONLY, NO_EVIDENCE, and INCONCLUSIVE_BASELINE_BUSY remain separate results with distinct exit codes. JSON output is available for automation.

    Honest limits

    The tool does not send messages or modify watched files. It classifies supplied observations and cannot prove that a file change was caused by a message. Terminal captures can contain private data and should be stored accordingly.

    What is inside

    • SKILL.md, README.md, LICENSE
    • scripts/delivery_evidence_judge.py
    • tests/test_delivery_evidence_judge.py
    • examples/ (the before/after fixture used in the example output)

    Requirements and safety

    Python 3.10+, standard library only. A terminal multiplexer (tmux) is needed only to capture a pane directly; saved pane text works without it. It makes no network calls.

    License

    Personal, non-transferable license for your own projects, including commercial work. No resale or redistribution.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.

    Classification is evidence-based, not causal. Snapshots may capture private terminal data. Requires manual capture before and after the event.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 5 days ago

    Software
    Python 3.10+
    Optional
    tmux (only to capture a pane directly)

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed5 days ago

    What's inside

    Frequently Asked Questions

    No. DELIVERED means two independent observations agree: the message appears as a new history line in the terminal capture, and at least one watched file changed by SHA-256. A changed hash is evidence of a content change, not proof that the message caused it.

    Any setup where you can capture the terminal text and name the files to watch. It reads a tmux pane (--target) or a saved text capture (--pane-text) and prints text or --json. It does not send messages itself.

    The judge reports INCONCLUSIVE_BASELINE_BUSY (exit code 4), because activity in the pane cannot be attributed to this send. A busy baseline is never counted as confirmed delivery.

    A ZIP with SKILL.md, README.md, LICENSE, scripts/delivery_evidence_judge.py (Python 3.10+, standard library only), tests/test_delivery_evidence_judge.py and examples/, the before/after fixture used in the example output. It is a personal license for your own projects, including commercial work.

    No. It never sends messages and never modifies the watched files. The only file it writes is the snapshot JSON you choose with --out. It makes no network calls. Snapshots contain the captured terminal text, so store them where private data may be kept.

    New in Testing & Debugging