Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Vulnerability Classes: From Bug Pattern to Primitive

    1

    Misjudge the class and you burn a week chasing a crash that was never going to give you a write.

    $9.99

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Vulnerability Classes: From Bug Pattern to Primitive

    Vulnerability Classes: From Bug Pattern to Primitive

    Example session with this skill installed

    I have a use-after-free in a C image parser (NotesApp's thumbnail decoder). Classify it, tell me the primitive, and what I'd need to turn it into something useful.

    • Read your context and instructions
    • Compiled the vulnerability classes

    Vulnerability class review: NotesApp thumbnail decoder

    Verdict (example run): Use-after-free on the heap, reachable from an untrusted image file. It points to an arbitrary read/write primitive once the heap is groomed.

    Classification

    FieldFinding
    ClassUse-after-free (temporal)
    EntryUntrusted thumbnail, attacker-controlled size
    Trust boundaryObject freed on a parse error, reused on the retry path
    CorruptionDangling pointer dereferenced after realloc
    PrimitiveControlled write into a freed chunk, then arbitrary read/write

    What stands between you and control

    ASLR and CFG are on, so a fixed address or a raw function-pointer overwrite will not fly. You need an info leak first, then a clean write target.

    Impact (illustrative)

    CVSS v4.0 8.4 (High), no privileges, triggered by a crafted file, in line with the UAF case study bundled in the skill.

    Next

    1. Confirm the free and reuse window with AddressSanitizer.
    2. Groom the heap to place a controllable object in the freed slot.
    3. Leak a pointer, then pivot the write into a read/write primitive.

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

    About this skill

    Most researchers can spot that something looks wrong, then stall on the question that matters: what can this bug actually do. Misjudge the class and you burn a week chasing a crash that was never going to give you a write. This skill makes your agent reason the way an exploit developer does, from the bug pattern to the primitive it really yields.

    What your agent does

    • Classifies the bug: stack or heap overflow, use-after-free, OOB read, uninitialized memory, race/TOCTOU, type confusion, integer wrap, or format string.
    • Traces the lifecycle: where untrusted input enters, how size, type and state are validated, and where corruption or a state desync happens.
    • Names the primitive: info leak, constrained write, arbitrary read/write, control-flow hijack, or auth bypass, not a vague ‘looks exploitable’.
    • Checks the mitigations: ASLR, DEP, /GS, CFG/CET, Safe-Linking and VBS, so you know whether chaining is even required.
    • Scores the real impact: CVSS v4.0 with EPSS and CISA KEV context, then a clean, reproducible PoC plan card.

    Built-in guard rails

    • Learning and research only: the examples are public CVE patterns and lab snippets, not weaponized payloads against a live target.
    • No primitive it can't prove: it will not claim an arbitrary write when the path only supports a leak.
    • Authorized use: it assumes you have permission for any target you apply it to.

    What's inside

    One SKILL.md of about 12,800 words across 9 sections, covering six vulnerability classes with worked CVE case studies, plus a LICENSE.txt.

    Works with Claude Code, Codex, Cursor, and any agent that loads SKILL.md.

    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.

    This is a learning and research reference, not an exploit generator. Reproducing the CVE case studies needs your own lab VM and debugger. It assumes you have authorization for any target you apply it to.

    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 1 day ago

    Software
    Claude Code
    orCodex
    orCursor
    Optional
    GDB
    orWinDbg
    lab VM for the CVE case studies

    Permissions

    Read Files
    Write Files
    Network Access
    Listed1 day ago

    What's inside

    Frequently Asked Questions

    No. It classifies the bug, names the primitive, and plans the PoC steps. You build and run the PoC in your own lab.

    No. The examples are public CVE patterns and lab snippets for learning, not weaponized payloads against a live target.

    An agent that loads SKILL.md. A debugger and a lab VM help if you want to reproduce the case studies.

    This one is memory-corruption and logic focused. For fast web and API triage, see the Fast Pentest Checklist skill.

    New in Security