Skip to content
Get packs

Zero-Day Discovery Engine

Guides an agent to find unknown vulnerabilities via variant analysis, fuzzing, taint tracing and N-day exploit rebuilding.

Source Download skill
Repository
AIPentest/CyberStrikeAI
Stars
8,323
Licence
Apache-2.0
Files
1 file

What it does

When public vulnerability databases come up empty, this skill tells the agent to stop matching known CVEs and instead reason about how code, protocols or APIs work and where they might break. It gives a structured method: find variants of past patches, look for gaps in patch filters, compare how two components parse the same input, fuzz with harnesses, trace untrusted data to dangerous sinks, and apply a checklist of trust-boundary and edge-case questions. It also covers rebuilding exploits from published patch diffs and sets a verification bar (reproducible PoC, root cause, proven impact, false positives excluded).

Great for

  • Security researchers doing authorized vulnerability research on their own or in-scope code
  • Red teams in sanctioned penetration tests who need a methodical discovery workflow
  • Students learning vulnerability research methods such as variant analysis and fuzzing

Skip it if

  • Attacking systems without written authorization, since the content is offensive and the skill does not check scope or permission
  • Non-security tasks, because the guidance is narrow and assumes familiarity with tools like CodeQL, AFL++ or bindiff
  • Expecting a turnkey scanner, since it is a methodology prompt and the agent still needs to run tools and reason about results

Try asking

  • “No public CVE matches this login endpoint; help me look for an unknown bug in its file upload handling.”
  • “Here is a security patch diff; find other places in this codebase with the same vulnerable pattern.”
  • “Write a CodeQL query to check whether user input in this Express app can reach exec or SQL sinks.”
  • “Set up a libFuzzer harness around this C parsing function and triage any crashes you find.”
Step-by-step guide: how to use Zero-Day Discovery EngineInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    zero-day-discovery.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip zero-day-discovery.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip zero-day-discovery.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Provide the target source code, patch diff or endpoint details up front, since the method relies on concrete artifacts.
  • Confirm written authorization and scope before any testing, and keep work in a lab or sanctioned environment.
  • Ask the agent to produce a minimal reproducible PoC and root-cause explanation, as the skill's verification standard requires.

What's inside

  • SKILL.md

Redistributed unmodified from AIPentest/CyberStrikeAI under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.