Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

CTF Binary Exploitation (Pwn) · ljagiello | DecimalAI

Securityby ljagiello

CTF Binary Exploitation (Pwn)

Provides binary exploitation techniques for CTF challenges.

Maintained

SkillScore

SkillScore70/100

Test results 70

Safety

3/3 safety checks passed

Code scan
Intent
Content

What it changes

61%70%of test cases passed +9 pts

Without skill14 of 23

With skill16 of 23

3 fixed13 passed either way6 still failing1 broke

Model 9/19/2026v2 · latest is v4

This benchmark measured an earlier version of the skill.

A newer attempt on gemini-3.8-flash on 10/24/2026 did not produce a publishable comparison (graded differently). The snapshot above is the previously measured run.

Track record

Agent runs: 0; All time

See one task, with and without the skill

One test taskIn a challenge where a global path flag sits right after a buffer and before a canary, an engineer wants to pad the input path with `A` characters to trigger a flag check. Why will `A` characters fail, and what path prefix character should be used as non-destructive padding?

Without the skillMissed

With the skillPassed

Same model, same task. Answers are capped and redacted.

Use this skill

Your agent runs on

Adds every file the skill has, pinned so you can update it. No account needed.

pip install "decimalai>=0.15"
decimalai skills add ljagiello-ctf-pwn
claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

The same install from Python, for a notebook or a setup script.

Install

pip install "decimalai>=0.15"

Add the skill

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

print("Added. Start Claude Code with the command below.")

Start Claude Code with it

claude --plugin-dir "$(git rev-parse --show-toplevel 2>/dev/null || pwd)/.agents"

Type these in a Claude Code session. No account or API key. It writes the skill and any files it bundles, and asks first when its safety rating isn't Passed.

Add the DecimalAI plugin, once

/plugin marketplace add decimal-labs/decimalai-claude-plugins
/plugin install decimalai@decimalai

Install this skill

/decimalai:install ljagiello-ctf-pwn

The plugin needs python3 and curl on your machine.

No Python or account. Claude Code picks it up in its next session.

Save it into your project

curl -fsSL --create-dirs -o .claude/skills/ljagiello-ctf-pwn/SKILL.md https://app.decimal.ai/s/ljagiello-ctf-pwn/SKILL.md

Copy SKILL.md

Save it in your project, as

.claude/skills/ljagiello-ctf-pwn/SKILL.md

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Claude Code docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

The same install from Python, for a notebook or a setup script. agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[claude-agent-sdk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from claude_agent_sdk import ClaudeAgentOptions
from decimalai.claude_agent_sdk import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

async def main():
    question = "Hello!"
    async with support.run(question) as run:  # one user message, one run
        async for message in run.query(options=ClaudeAgentOptions(max_turns=6)):
            print(message)

# In a notebook or async code: await main()
asyncio.run(main())

Prefer its own loader? ClaudeAgentOptions(plugins=[skills_plugin()], skills="all") reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from claude_agent_sdk import ClaudeAgentOptions, query

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
options = ClaudeAgentOptions(system_prompt=f"You are a helpful agent.\n\n{skill}", max_turns=6)

async def main():
    async for message in query(prompt="Hello!", options=options):
        print(message)

# In a notebook or async code: await main()
asyncio.run(main())

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Claude Agent SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai-agents]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

Install

pip install "decimalai[openai-agents]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from agents import Agent, Runner
from decimalai.openai_agents import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    name=support.name,
    instructions=support.instructions,  # your prompt, then the skill list
    tools=[*support.tools],             # your tools, then load_skill
)
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

Using SandboxAgent? capabilities=[Shell(), decimalai.openai_agents.sandbox_skills()] points its own loader at the same folder.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from agents import Agent, Runner

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = Agent(name="support", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await Runner.run(agent, "Hello!")
result = Runner.run_sync(agent, "Hello!")

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

OpenAI Agents docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[openai]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

Install

pip install "decimalai[openai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import json

from openai import OpenAI
from decimalai.openai import agent_config

client = OpenAI()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your tools, as Chat Completions takes them, and what answers each.
TOOLS = [{"type": "function", "function": {
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters": {"type": "object", "properties": {"order_id": {"type": "string"}},
                   "required": ["order_id"]},
}}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    # run.system is your prompt, then the skill list.
    messages = [{"role": "system", "content": run.system}, run.message]
    while True:
        reply = client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=messages,
            tools=[*TOOLS, *run.tools],  # your tools, then load_skill
        ).choices[0].message
        messages.append(reply)
        if not reply.tool_calls:
            break
        for call in reply.tool_calls:
            # load_skill is answered by the run; your own tools by FUNCTIONS.
            function = FUNCTIONS.get(call.function.name)
            messages.append(run.tool_result(call) or {
                "role": "tool", "tool_call_id": call.id,
                "content": function(**json.loads(call.function.arguments)) if function
                else f"There is no tool named {call.function.name!r}.",
            })
print(reply.content)

# The Responses API takes run.response_tools instead of run.tools.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from openai import OpenAI

client = OpenAI()
skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
reply = client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[
        {"role": "system", "content": f"You are a helpful agent.\n\n{skill}"},
        {"role": "user", "content": "Hello!"},
    ],
).choices[0].message

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

OpenAI SDK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[anthropic]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

Install

pip install "decimalai[anthropic]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import anthropic
from decimalai.anthropic import agent_config

client = anthropic.Anthropic()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

question = "Hello!"
with support.run(question) as run:  # one user message, one run
    answer = client.beta.messages.tool_runner(
        model="claude-haiku-5-5",
        max_tokens=4096,
        system=run.system,     # your prompt, then the skill list
        tools=[*run.tools],    # your tools, then load_skill
        messages=[{"role": "user", "content": question}],
    ).until_done()

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
import anthropic

client = anthropic.Anthropic()
skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
reply = client.messages.create(
    model="claude-haiku-5-5",
    max_tokens=4096,
    system=f"You are a helpful agent.\n\n{skill}",
    messages=[{"role": "user", "content": "Hello!"}],
)

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Anthropic docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[google-genai]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

Install

pip install "decimalai[google-genai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from google import genai
from google.genai import types
from decimalai.google_genai import agent_config

client = genai.Client()
# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")


def lookup_order(order_id: str) -> str:
    return f"Order {order_id} shipped yesterday."


# Your functions, as declarations for the model, and what answers each.
TOOLS = [{"function_declarations": [{
    "name": "lookup_order",
    "description": "Look up the status of an order.",
    "parameters_json_schema": {"type": "object", "properties": {"order_id": {"type": "string"}},
                               "required": ["order_id"]},
}]}]
FUNCTIONS = {"lookup_order": lookup_order}

question = "Where is order A1001?"
with support.run(question) as run:  # one user message, one run
    config = types.GenerateContentConfig(
        system_instruction=run.system,     # your prompt, then the skill list
        tools=[*TOOLS, *run.tool_params],  # your functions, then load_skill
    )
    contents = [{"role": "user", "parts": [{"text": run.message}]}]
    while True:
        response = client.models.generate_content(
            model="gemini-3.5-flash", contents=contents, config=config,
        )
        if not response.function_calls:
            break
        contents.append(response.candidates[0].content)
        parts = []
        for call in response.function_calls:
            # load_skill is answered by the run; your own functions by FUNCTIONS.
            # A loop of your own rather than automatic function calling, which
            # raises KeyError when the model asks for a function you did not pass.
            function = FUNCTIONS.get(call.name)
            parts.append(run.tool_result(call) or {"function_response": {
                "name": call.name, "id": call.id,
                "response": {"result": function(**(call.args or {})) if function
                             else f"There is no function named {call.name!r}."},
            }})
        contents.append({"role": "user", "parts": parts})
print(response.text)

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from google import genai
from google.genai import types

client = genai.Client()
skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
response = client.models.generate_content(
    model="gemini-3.5-flash",
    contents="Hello!",
    config=types.GenerateContentConfig(system_instruction=f"You are a helpful agent.\n\n{skill}"),
)

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Gemini API docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[langchain]" langchain-openai
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

Install

pip install "decimalai[langchain]" langchain-openai

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from langchain.agents import create_agent
from decimalai.langchain import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = create_agent(
    "openai:gpt-5.4-mini",
    tools=[],                         # your tools; the middleware adds load_skill
    middleware=[support.middleware],  # your prompt, the skill list and load_skill
)
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from langchain.agents import create_agent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = create_agent("openai:gpt-5.4-mini", tools=[], system_prompt=f"You are a helpful agent.\n\n{skill}")
result = agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

LangChain docs

CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

The same install from Python, for a notebook or a setup script. CrewAI's own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" "crewai>=1.15.27"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.crewai
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                               # your own CrewAI LLM
    skills=decimalai.crewai.skills(),      # the installed folder
    tools=decimalai.crewai.skill_tools(),  # reads a skill's files, runs its scripts
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

No DecimalAI package. CrewAI's own loader reads the folder, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from crewai import Agent, Crew, Task

agent = Agent(
    role="Support", goal="Help the customer", backstory="A support agent.",
    llm=llm,                         # your own CrewAI LLM
    skills=[Path(".agents/skills")],
)
task = Task(description="Hello!", expected_output="an answer", agent=agent)
Crew(agents=[agent], tasks=[task]).kickoff()

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

CrewAI docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[adk]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

Install

pip install "decimalai[adk]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.adk import DecimalAIPlugin, agent_config
from google.adk.agents import LlmAgent
from google.adk.apps import App

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = LlmAgent(
    name="support",          # ADK needs a Python identifier
    model="gemini-3.5-flash",
    instruction=support.instruction,  # your prompt, then the skill list
    tools=[*support.tools],           # your tools, then load_skill
)
# DecimalAIPlugin makes each invocation of runners built from this App one trace.
app = App(name="support", root_agent=agent, plugins=[DecimalAIPlugin()])

Prefer ADK's own loader? SkillToolset(skills=decimalai.adk.skills(), ...) reads the same folder. Use one or the other: a skill delivered both ways reaches the prompt twice.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from google.adk.agents import LlmAgent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = LlmAgent(
    name="support",
    model="gemini-3.5-flash",
    instruction=lambda context: f"You are a helpful agent.\n\n{skill}",
)

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Google ADK docs

agent_config() hands your agent the skill. With no DecimalAI key it reads your project's files and sends nothing. Needs SDK 0.27 or newer.

Install

pip install "decimalai[pydantic-ai]"
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

Install

pip install "decimalai[pydantic-ai]"

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from pydantic_ai import Agent
from decimalai.pydantic_ai import agent_config

# With no DECIMAL_API_KEY this reads the skills your project lists, from disk.
support = agent_config("support", fallback="You are a helpful agent.")

agent = Agent(
    "openai:gpt-4o-mini",
    name=support.name,
    tools=[],                 # your tools; the capability adds load_skill
    capabilities=[support],   # your prompt, the skill list and load_skill
)
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

On the Pydantic AI harness, decimalai.pydantic_ai_harness.skills(project) uses its own loader instead.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from pydantic_ai import Agent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = Agent("openai:gpt-4o-mini", instructions=f"You are a helpful agent.\n\n{skill}")
# In a notebook or async code: result = await agent.run("Hello!")
result = agent.run_sync("Hello!")

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Pydantic AI docs

Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

The same install from Python, for a notebook or a setup script. Deep Agents' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" deepagents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import decimalai.deepagents
from deepagents import create_deep_agent

agent = create_deep_agent(model=model, **decimalai.deepagents.skills())
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

The default backend reads files but has no shell. To run a skill's scripts, pass skills(backend=LocalShellBackend(...)).

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from deepagents import create_deep_agent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = create_deep_agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent.invoke({"messages": [{"role": "user", "content": "Hello!"}]})

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Deep Agents docs

Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

The same install from Python, for a notebook or a setup script. Its own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" agent-framework

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

import asyncio

from agent_framework import Agent
from decimalai.agent_framework import skills

agent = Agent(client=client, instructions="You are a helpful agent.", context_providers=[skills()])
asyncio.run(agent.run("Hello!"))

It caches skills by default, so restart the process after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

import asyncio
from pathlib import Path
from agent_framework import Agent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = Agent(client=client, instructions=f"You are a helpful agent.\n\n{skill}")
asyncio.run(agent.run("Hello!"))

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Microsoft Agent Framework docs

Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents
decimalai skills add ljagiello-ctf-pwn

Add it to your agent

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

The same install from Python, for a notebook or a setup script. Strands' own skill loader reads the installed folder. No DecimalAI account needed.

Install

pip install "decimalai>=0.16" strands-agents

Add it to your agent

from decimalai import skillpkg

# Writes .agents/skills and .decimal/skills.toml. Safe to run twice.
result = skillpkg.add("ljagiello-ctf-pwn")
if result.exit_code:
    raise SystemExit(result.problems)

from decimalai.strands import skills
from strands import Agent
from strands.vended_tools import file_editor, shell

agent = Agent(
    model=model,                # always pass one: the default is Amazon Bedrock
    system_prompt="You are a helpful agent.",
    tools=[file_editor, shell],
    plugins=[skills()],
)
agent("Hello!")

Strands loads an agent's skills once, so build the agent again after decimalai skills update.

No DecimalAI package. The whole skill goes in every prompt, and nothing pins or updates it.

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

Give it to your agent

from pathlib import Path
from strands import Agent

skill = Path(".agents/skills/ljagiello-ctf-pwn/SKILL.md").read_text()
agent = Agent(model=model, system_prompt=f"You are a helpful agent.\n\n{skill}")
agent("Hello!")

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Strands docs

Adds every file the skill has, pinned so you can update it. Cursor, Copilot, Cline and Warp read this folder by themselves.

pip install "decimalai>=0.15"
decimalai skills add ljagiello-ctf-pwn

No Python or account. Cursor, Copilot, Cline and Warp read this folder by themselves.

Save it into your project

curl -fsSL --create-dirs -o .agents/skills/ljagiello-ctf-pwn/SKILL.md https://app.decimal.ai/s/ljagiello-ctf-pwn/SKILL.md

Copy SKILL.md

Save it in your project, as

.agents/skills/ljagiello-ctf-pwn/SKILL.md

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Cursor and other editors docs

Fetch the file and give its text to your model as instructions. No account or package needed.

Get the file

curl -fsSL https://app.decimal.ai/s/ljagiello-ctf-pwn/SKILL.md

Read the summary as JSON

curl -fsSL https://app.decimal.ai/s/ljagiello-ctf-pwn.json

This gets SKILL.md only. The skill's 5 other files are not included.

Other ways to add it

Other setups docs

You don't need a DecimalAI account.

Not for

The main blocker is understanding what the binary does; use reverse engineering first. D

Version: v4 (v2 tested 9/19/2026)
Source: ljagiello/ctf-skills ↗
License: MIT

Manage it in DecimalAI

Saved prompts and skills, pinned versions, and traces of this skill in your runs. DecimalAI does not run your agent.

Claude Code reads files from your project, so DecimalAI config does not apply there.

On Claude Agent SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI Agents: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On OpenAI SDK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Anthropic: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Gemini API: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On LangChain: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On CrewAI: set DECIMAL_API_KEY and add decimalai.init(crewai=True) to record runs and skill use. The skill version stays in your repo.

On Google ADK: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Pydantic AI: save the skill to an agent here, set DECIMAL_API_KEY, and the same code uses it with no command to run. Changes you save reach the next run with no redeploy.

On Deep Agents: set DECIMAL_API_KEY and add decimalai.init(langchain=True) to record runs and skill use. The skill version stays in your repo.

On Microsoft Agent Framework: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

On Strands: set DECIMAL_API_KEY and add decimalai.init(otel=True, agent_name="support") to record runs and skill use. The skill version stays in your repo.

Editors read files from your project, so DecimalAI config does not apply there.

Anywhere else: record a use with decimalai.log_skill_activation() inside a trace.

Use this skill

SKILL.md

CTF Binary Exploitation (Pwn)

Quick reference for binary exploitation (pwn) CTF challenges. Each technique has a one-liner here; see supporting files for full details.

Prerequisites

Python packages (all platforms):

pip install "pwntools==4.15.0" "ROPgadget==7.7" "ropper==1.13.13"
# or via script: bash scripts/install_ctf_tools.sh python

uv alternative:

uv venv && uv pip install "pwntools==4.15.0" "ROPgadget==7.7" "ropper==1.13.13"

Linux (apt):

apt install gdb binutils strace ltrace qemu-system-x86

Optional cross-arch (ARM/MIPS — Debian/Ubuntu):

apt install qemu-user qemu-user-static gdb-multiarch binutils-multiarch libc6-armhf-cross libc6-arm64-cross libc6-mips-cross

macOS (Homebrew):

brew install gdb binutils qemu

Ruby gems (all platforms):

gem install one_gadget seccomp-tools

Manual install:

pwndbg — Linux: GitHub, macOS: brew install pwndbg/tap/pwndbg-gdb
checksec — included with pwntools (checksec --file=binary)

MCP servers (optional): WinDbg MCP (svnscha/mcp-windbg) for Windows crash dumps, live breakpoint-driven exploitation, and kernel/driver debugging; IDA Pro MCP / Ghidra MCP for decompiling the vulnerable function — see tools-mcp.md.

Additional Resources

overflow-basics.md - Stack/global buffer overflow, ret2win, canary bypass, canary byte-by-byte brute force on forking servers, struct pointer overwrite, signed integer bypass, hidden gadgets, stride-based OOB read leak, parser stack overflow via unchecked memcpy length with callee-saved register restoration
rop-and-shellcode.md - Core ROP chains (ret2libc, syscall ROP, rdx control, shell interaction), ret2csu, bad character XOR bypass, exotic x86 gadgets (BEXTR/XLAT/STOSB/PEXT), stack pivot via xchg rax,esp, sprintf() gadget chaining for bad character bypass, canary XOR epilogue as RDX zeroing gadget, stub_execveat syscall as execve alternative via read() return value
rop-advanced.md - Advanced ROP techniques: double stack pivot to BSS via leave;ret, SROP (Sigreturn-Oriented Programming) with UTF-8 constraints, seccomp bypass, RETF architecture switch (x64→x32) for seccomp bypass, shellcode with input reversal, .fini_array hijack, ret2vdso, pwntools template, x32 ABI syscall aliasing for seccomp bypass, time-based blind shellcode exfiltration
format-string.md - Format string exploitation (leaks, GOT overwrite, blind pwn, filter bypass, canary leak, __free_hook, .rela.plt patching, saved EBP overwrite for .bss pivot, argv[0] overwrite for stack smash info leak, .fini_array loop for multi-stage exploitation, __printf_chk bypass with sequential %p, single-call leak + GOT overwrite, ROT13-encoded format string exploit through input transformation)
advanced.md - Seccomp advanced techniques, UAF, JIT, esoteric GOT, heap overlap via base conversion, tree data structure stack underallocation, ret2dlresolve, kernel exploitation (basic)
heap-techniques.md - House of Apple 2 (+ setcontext SUID variant), House of Einherjar, House of Orange/Spirit/Lore/Force, heap grooming, custom allocators (nginx, talloc), classic unlink, musl libc heap (meta pointer + atexit hijack), tcache stashing unlink attack, unsafe unlink + top chunk consolidation
heap-techniques-2.md - CTF-writeup heap variants: UAF vtable pointer encoding shell argument, uninitialized chunk residue pointer leak, tcache strcpy null-byte overflow + backward consolidation, adjacent-struct fn-pointer overflow for libc leak + GOT overwrite, hidden-menu tcache poisoning, tcache double-free + fake _IO_FILE vtable stdout hijack, tcache-to-fastbin promotion cross-bin attack, 6-bit index OOB + written_bytes accumulator, IS_MMAPED bit-flip for unsorted bin leak on calloc'd chunk, filename-regex-constrained fastbin via LSB-only heap pointer overwrite, custom allocator unsafe unlink to GOT
heap-fsop.md - FILE-structure (_IO_FILE) exploitation: fastbin stdout vtable two-stage hijack for PIE + Full RELRO, _IO_buf_base null-byte stdin hijack, glibc 2.24+ _IO_FILE vtable validation bypass, unsorted-bin attack on stdin IO_buf_end, unsorted-bin corruption via mp structure, realloc(ptr, 0) as free() UAF, single-byte reference counter wraparound UAF
advanced-exploits.md - Advanced exploit techniques (part 1): VM signed comparison, BF JIT shellcode, type confusion, off-by-one index corruption, DNS overflow, ASAN shadow memory, format string with encoding constraints, custom canary preservation, signed integer bypass, canary-aware partial overflow, CSV injection, MD5 preimage gadgets, VM GC UAF slab reuse, path traversal sanitizer bypass, FSOP + seccomp bypass via openat/mmap/write
advanced-exploits-2.md - Advanced exploit techniques (part 2): bytecode validator bypass via self-modification, io_uring UAF with SQE injection, integer truncation int32->int16, GC null-reference cascading corruption, leakless libc via multi-fgets stdout FILE overwrite, signed/unsigned char underflow heap overflow, XOR keystream brute-force write primitive, tcache pointer decryption heap leak, unsorted bin promotion via forged chunk size, FSOP stdout TLS leak, TLS destructor hijack via __call_tls_dtors, custom shadow stack pointer overflow bypass, signed int overflow negative OOB heap write, XSS-to-binary pwn bridge
advanced-exploits-4.md - Advanced exploit techniques (part 4): Windows SEH overwrite + pushad VirtualAlloc ROP, IAT-relative resolution, detached process shell stability, SeDebugPrivilege SYSTEM escalation, ARM buffer overflow with Thumb shellcode, Forth interpreter system word exploitation, GF(2) Gaussian elimination for multi-pass tcache poisoning, single-bit-flip exploitation primitive (mprotect + iterative code patching), Game of Life shellcode evolution via still-lifes, UAF via menu-driven strdup/free ordering, Windows CFG bypass via system() as valid call target, neural network output as function pointer index OOB, shellcode unique-byte limit bypass via counter overflow
advanced-exploits-3.md - Advanced exploit techniques (part 3): stack variable overlap / carry corruption OOB, 1-byte overflow via 8-bit loop counter, game AI arithmetic mean OOB read, arbitrary read/write GOT overwrite to shell, stack leak via __environ + memcpy overflow, JIT sandbox escape via uint16 jump truncation, DNS compression pointer stack overflow with multi-question ROP, ELF code signing bypass via program header manipulation, game level format signed/unsigned coordinate mismatch, file descriptor inheritance via missing O_CLOEXEC, sign extension integer underflow in metadata parsing, ROP chain construction with read-only primitive, 4-byte shellcode with timing side-channel via persistent registers, CRC oracle as arbitrary read, UTF-8 case conversion buffer overflow
advanced-exploits-5.md - Advanced exploit techniques (part 5): data-interpretation exploitation — Chip-8 emulator OOB memory for ret2libc, double-precision float quicksort canary repositioning, bloom filter abs(INT_MIN) negative index OOB write
sandbox-escape.md - Custom VM exploitation, FUSE/CUSE devices, busybox/restricted shell, shell tricks, process_vm_readv sandbox bypass, named pipe file size bypass, CPU emulator print opcode Python eval injection (cross-references ctf-misc/pyjails.md for Python jail techniques)
kernel.md - Linux kernel exploitation fundamentals: environment setup, QEMU debug, heap spray structures (tty_struct, poll_list, user_key_payload, seq_operations), kernel stack overflow, canary leak, privilege escalation (ret2usr, kernel ROP), modprobe_path overwrite, core_pattern overwrite, kmalloc size mismatch heap overflow + struct file f_op corruption
kernel-techniques.md - Kernel exploitation techniques: tty_struct kROP (fake vtable + stack pivot), AAW via ioctl register control, userfaultfd race stabilization, SLUB allocator internals (freelist hardening/obfuscation), leak via kernel panic, MADV_DONTNEED race window extension (DiceCTF 2026), cross-cache CPU-split attack (DiceCTF 2026), PTE overlap file write (DiceCTF 2026), addr_limit bypass via failed file open for kernel memory read/write
kernel-bypass.md - Kernel protection bypass: KASLR/FGKASLR bypass (__ksymtab), KPTI bypass (swapgs trampoline, signal handler, modprobe_path/core_pattern via ROP), SMEP/SMAP bypass, GDB kernel module debugging, initramfs/virtio-9p workflow, exploit templates, exploit delivery
field-notes.md - Detailed pwn notes: heap exploitation quick reference, additional exploit notes, useful commands
attack-playbook.md - Linux userland attack playbook: recon of input/privilege boundaries, leak selection by mitigation, and primitive ranking workflow
static-signals.md - Exploit-relevant static signals (libc version, seccomp policy, RELRO/PIE/canary/NX evidence) for exploitability assessment
pwnable-tw/README.md - 43 pwnable.tw writeups (orw, start, calc, hacknote, babystack, tcache_tear, death_note, dubblesort, applestore, bookwriter, and more) with per-challenge environment, protections, root-cause defects, and concrete exploitation chains
tools-mcp.md - MCP server integration for pwn: WinDbg MCP session model (open_cdb_dump/open_cdb_remote/open_kd_session + session_id), crash-dump triage, in-dump flag{ memory scanning, live breakpoint/hardware-watchpoint/memory-patch exploitation, !dh -f Windows mitigation check (the checksec equivalent), kernel-mode token stealing with kd; IDA/Ghidra MCP for exploit development (decompile the vulnerable function first, drop to assembly only for exact gadget bytes)

When to Pivot

If you do not yet understand what the binary does, switch to /ctf-reverse before trying to exploit it.
If the service is really a restricted shell, encoding puzzle, or sandbox language challenge, switch to /ctf-misc.
If the exploit path depends on a web endpoint, session bug, or upload primitive more than memory corruption, switch to /ctf-web.
If the vulnerability requires breaking a cryptographic primitive before exploitation, switch to /ctf-crypto.

Quick Start Commands

# Binary analysis
checksec --file=binary
file binary
readelf -h binary

# Find gadgets
ROPgadget --binary binary | grep "pop rdi"
ropper -f binary --search "pop rdi"
one_gadget /lib/x86_64-linux-gnu/libc.so.6

# Debug
gdb -q binary -ex 'start' -ex 'checksec'

# Pattern for offset finding
python3 -c "from pwn import *; print(cyclic(200))"
python3 -c "from pwn import *; print(cyclic_find(0x61616168))"

# libc identification
./libc-database/find puts <leaked_addr_last_3_nibbles>

Source Code Red Flags

Threading/pthread -> race conditions
usleep()/sleep() -> timing windows
Global variables in multiple threads -> TOCTOU

Race Condition Exploitation

bash -c '{ echo "cmd1"; echo "cmd2"; sleep 1; } | nc host port'

Common Vulnerabilities

Buffer overflow: gets(), scanf("%s"), strcpy()
Format string: printf(user_input)
Integer overflow, UAF, race conditions

Protection Implications for Exploit Strategy

Protection	Status	Implication
PIE	Disabled	All addresses (GOT, PLT, functions) are fixed - direct overwrites work
RELRO	Partial	GOT is writable - GOT overwrite attacks possible
RELRO	Full	GOT is read-only - need alternative targets (hooks, vtables, return addr)
NX	Enabled	Can't execute shellcode on stack/heap - use ROP or ret2win
Canary	Present	Stack smash detected - need leak or avoid stack overflow (use heap)

Quick decision tree:

Partial RELRO + No PIE -> GOT overwrite (easiest, use fixed addresses)
Full RELRO -> target __free_hook, __malloc_hook (glibc < 2.34), or return addresses
Stack canary present -> prefer heap-based attacks or leak canary first

Stack Buffer Overflow

Find offset: cyclic 200 then cyclic -l <value>
Check protections: checksec --file=binary
No PIE + No canary = direct ROP
Canary leak via format string or partial overwrite
Canary brute-force byte-by-byte on forking servers (7*256 attempts max)

ret2win with magic value: Overflow -> ret (alignment) -> pop rdi; ret -> magic -> win(). Stack alignment: SIGSEGV in movaps = add extra ret gadget. Offset: buffer at rbp - N, return at rbp + 8, total = N + 8. Input filtering: assert payload avoids memmem() banned strings. Gadgets: ROPgadget --binary binary | grep "pop rdi", or pwntools ROP() for hidden gadgets in CMP immediates. See overflow-basics.md for full exploit code.

Parser Stack Overflow (Unchecked memcpy)

Pattern: Custom file parser (PCAP, image, archive) allocates fixed stack buffer but input records can exceed it. memcpy copies before length validation, overflowing saved registers and return address. Must restore callee-saved registers: rbx to readable memory (BSS), loop counters to exit values, then ret gadget + win function. See overflow-basics.md.

Struct Pointer Overwrite (Heap Menu Challenges)

Pattern: Menu create/modify/delete on structs with data buffer + pointer. Overflow name into pointer field with GOT address, then write win address via modify. See overflow-basics.md for full exploit and GOT target selection table.

Signed Integer Bypass

Pattern: scanf("%d") without sign check; negative quantity * price = negative total, bypasses balance check. See overflow-basics.md.

Canary-Aware Partial Overflow

Pattern: Overflow valid flag between buffer and canary. Use ./ as no-op path padding for precise length. See overflow-basics.md and advanced.md for full exploit chain.

Global Buffer Overflow (CSV Injection)

Pattern: Adjacent global variables; overflow via extra CSV delimiters changes filename pointer. See overflow-basics.md and advanced.md for full exploit.

ROP Chain Building

Leak libc via puts@PLT(puts@GOT), return to vuln, stage 2 with system("/bin/sh"). See rop-and-shellcode.md for full two-stage ret2libc pattern, leak parsing, and return target selection.

DynELF libc discovery: pwntools.DynELF(leak_func, pointer_in_libc) resolves libc symbols remotely without knowing the libc version. See rop-and-shellcode.md.

Constrained shellcode in small buffers: When buffer is too small, use read() shellcode stub (< 20 bytes) to pull full stage-2 shellcode. See rop-and-shellcode.md.

Raw syscall ROP: When system()/execve() crash (CET/IBT), use pop rax; ret + syscall; ret from libc. See rop-and-shellcode.md.

ret2csu: __libc_csu_init gadgets control rdx, rsi, edi and call any GOT function — universal 3-argument call without libc gadgets. See rop-and-shellcode.md.

Bad char XOR bypass: XOR payload data with key before writing to .data, then XOR back in place with ROP gadgets. Avoids null bytes, newlines, and other filtered characters. See rop-and-shellcode.md.

Exotic gadgets (BEXTR/XLAT/STOSB/PEXT): When standard mov write gadgets are unavailable, chain obscure x86 instructions for byte-by-byte memory writes. See rop-and-shellcode.md.

Stack pivot (xchg rax,esp): Swap stack pointer to attacker-controlled heap/buffer when overflow is too small for full ROP chain. Requires pop rax; ret to load pivot address first. See rop-and-shellcode.md.

rdx control: After puts(), rdx is clobbered to 1. Use pop rdx; pop rbx; ret from libc, or re-enter binary's read setup + stack pivot. See rop-and-shellcode.md.

Canary XOR epilogue as rdx zeroing gadget: When no pop rdx; ret exists, jump into the canary check epilogue xor rdx, fs:28h -- it zeros RDX when the canary is intact. See rop-and-shellcode.md.

stub_execveat as execve alternative: When no pop rax; ret exists, use stub_execveat (syscall 322/0x142) instead of execve -- send exactly 0x142 bytes so read() return value sets rax. See rop-and-shellcode.md.

Shell interaction: After execve, sleep(1) then sendline(b'cat /flag*'). See rop-and-shellcode.md.

Format String Through Input Transformation

ROT13-encoded format string: When input is ROT13/Caesar-transformed before reaching printf, pre-encode the format string payload with the inverse transform so it arrives intact. See format-string.md.

Kernel Exploitation

addr_limit bypass via failed file open: When a kernel module sets addr_limit = KERNEL_DS but fails to restore it on error paths, force the error (e.g., make target file a directory) to retain kernel memory access from userspace read()/write(). See kernel-techniques.md.

Sandbox and Emulator Escape

CPU emulator eval injection: When an emulator's print opcode uses eval('"' + buf + '"') for escape sequences, build "+__import__("os").system("cmd")# in emulator memory via ADD opcodes to escape the string and execute Python. See sandbox-escape.md.

Advanced Exploit Primitives

Neural network function pointer OOB: When a binary uses NN output as an index into a function pointer array without bounds checking, retrain weights/biases to produce an out-of-bounds index that reads a target address from the biases array. See advanced-exploits-4.md.

Shellcode unique-byte limit bypass via counter overflow: When shellcode is limited to N unique bytes, spray the stack to corrupt the seen[256] counter, then re-execute main (skipping memset) so the overflowed counter allows arbitrary bytes on the second run. See advanced-exploits-4.md.

Deep-Dive Notes

Use field-notes.md once you have confirmed the challenge is truly exploitation-heavy.

Heap and allocator notes: House of Apple, tcache, unsafe unlink, talloc, UAF, FSOP
Advanced exploit notes: seccomp bypass, ret2vdso, io_uring, integer truncation, ASAN, timing oracles
Sandbox and hybrid notes: pyjail crossover, busybox escapes, custom VMs, shell tricks, path sanitizers
Kernel and Windows notes: kernel playbooks, SEH, CFG bypass, privilege escalation
Historical case notes: older but still reusable CTF exploit patterns

When it activates

Automatically, when a task matches. Use when you already have a vulnerable native target or service and need to turn memory corruption or low-level primitives into code execution or privilege escalation, such as buffer overflows, format strings, heap bugs, ROP, ret2libc, shellcode, kernel exploitation, seccomp bypass, sandbox escape, or Windows/Linux exploit chains.

About this skill

Automatic

Security & AppSec pack

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v8 — 10/23/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

ljagiello / ctf-skills

ctf-pwn/SKILL.md

View on GitHub ↗

License MITImported 10/31/2026Version v4Bundled 5 files · scripts

Illustrative Want to see what ljagiello/ctf-pwn would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Related skills

Skills that do similar work, with their headline benchmark lift.

From the same source and nearby in the catalogue.

Ready to supercharge your AI agents?

SkillScore

Safety

What it changes

Track record

Use this skill

Not for

CTF Binary Exploitation (Pwn)

Prerequisites

Additional Resources

When to Pivot

Quick Start Commands

Source Code Red Flags

Race Condition Exploitation

Common Vulnerabilities

Protection Implications for Exploit Strategy

Stack Buffer Overflow

Parser Stack Overflow (Unchecked memcpy)

Struct Pointer Overwrite (Heap Menu Challenges)

Signed Integer Bypass

Canary-Aware Partial Overflow

Global Buffer Overflow (CSV Injection)

ROP Chain Building

Format String Through Input Transformation

Kernel Exploitation

Sandbox and Emulator Escape

Advanced Exploit Primitives

Deep-Dive Notes

When it activates

SkillSafety

Source

Related skills

More in the registry