More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Case Review: Investigation Handoff Audit

    1

    Audit a finished investigation before it leaves your hands: every finding traced to evidence, confidence labeled per finding, language calibrated down, and a handoff a peer can reproduce.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Case Review: Investigation Handoff Audit

    Case Review: Investigation Handoff Audit

    Example session with this skill installed

    Our malware triage case is ready for handoff. The draft claims the sample beacons out and exfiltrates documents. Audit the package before it goes to the client.

    • Read your context and instructions
    • Compiled the case review

    INTAKE: All five pieces present — scope record (one host, one sample, two days), timeline, evidence inventory (12 artifacts), work log, draft report. Proceed.

    TRACEABILITY: Nine findings. Seven solid: evidence cited, corroborated where stakes are high. One thin: the exfiltration claim rests on a single network capture with no second observation. One broken: a persistence finding cites a registry key the inventory shows was never collected — it was copied from a different case's notes.

    RED-FLAG SWEEP: Two flags — the broken finding is also a stale artifact (the cited key belongs to another case), and the work log shows two failed hypothesis tests that never made it into the draft (missing negative results).

    CALIBRATION: "The malware proves the attacker targets finance" becomes "the observed activity is consistent with targeting finance-sector documents; attribution to a specific actor is not established." "Always beacons on startup" becomes "beaconed on startup in all 14 observed runs."

    OUTCOME: Handoff blocked. The persistence finding is removed pending re-collection, the exfiltration claim is downgraded to medium confidence and flagged single-source, and the two failed tests are added to the report. It ships a day later with a report the analyst can defend in front of a hostile reader.

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

    About this skill

    An investigation is not done when you know the answer — it is done when someone else can follow your trail, check your work, and defend your conclusions. This skill is the audit you run on your own case before anyone else sees it: intake first (scope record, timeline, evidence inventory, work log — if any is missing, the audit stops there), then a traceability check of every evidence-to-finding-to-conclusion chain, then a red-flag sweep that blocks the handoff (circular evidence, missing negative results, unverifiable steps, scope drift, stale artifacts).

    Language gets calibrated, not inflated: "proves" becomes "is consistent with," "the attacker" becomes "the observed activity," "always" becomes "in all 14 observed runs." The handoff package has five parts — executive summary, findings with evidence links, methodology, limitations, raw artifacts with hashes and chain of custody — and sign-off requires written "yes" answers to three reviewer questions, including whether you would defend the report in front of someone hostile to its conclusions. A worked example blocks a malware handoff over a copied-across finding and a single-source exfiltration claim, and ships it a day later defensible.

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

    • It audits the package, not the investigation — evidence that was never collected cannot be conjured by the audit.
    • A genuinely single-source case can only be labeled honestly, not corroborated by assertion.
    • Scope drift already committed must be disclosed, not erased.
    • Time pressure does not change the gate: a blocked handoff is the audit working, not a delay.

    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 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    No. The audit stops at intake. A work log that includes dead ends is part of the evidence — it shows what was ruled out. Fix the missing piece first, then audit.

    Thin means it cites evidence but rests on a single source where the stakes are high — it can ship with honest confidence labeling. Broken means it cites no evidence, or the cited evidence does not support it — that blocks the handoff.

    Because careful language signals careful work. A report gets more convincing as its claims get more precise — "consistent with" backed by evidence beats "proves" backed by enthusiasm, especially in front of a hostile reader.

    New in Security