competition-prompt-injection
Internal downstream skill for ctf-sandbox-orchestrator. CTF-sandbox workflow for prompt-injection, retrieval poisoning, memory contamination, planner drift, MCP or tool-boundary abuse, and agent exfiltration challenges. Use when the user asks to analyze prompt injection, retrieval poisoning, memory contamination, planner drift, tool-argument corruption, or secret exposure caused by an agent chain. Use only after `$ctf-sandbox-orchestrator` has already established sandbox assumptions and routed here.
- License 许可
- GPL-3.0
- 最近更新
- 2026-09-19
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- competition-prompt-injection
- description
- Internal downstream skill for ctf-sandbox-orchestrator. CTF-sandbox workflow for prompt-injection, retrieval poisoning, memory contamination, planner drift, MCP or tool-boundary abuse, and agent exfiltration challenges. Use when the user asks to analyze prompt injection, retrieval poisoning, memory contamination, planner drift, tool-argument corruption, or secret exposure caused by an agent chain. Use only after `$ctf-sandbox-orchestrator` has already established sandbox assumptions and routed here.
02 资源文件 2
- 参考文档agents/openai.yaml366 B
- 参考文档references/prompt-injection.md996 B
03 SKILL.md 正文
Competition Prompt Injection
Use this skill only as a downstream specialization after $ctf-sandbox-orchestrator is already active and has established sandbox assumptions, node ownership, and evidence priorities. If that has not happened yet, return to $ctf-sandbox-orchestrator first.
Use this skill when the challenge is primarily about trust boundaries inside an agentic system.
Reply in Simplified Chinese unless the user explicitly requests English.
Quick Start
- Identify the first untrusted content that becomes model-visible.
- Map the chain from retrieval, memory, or transcript into planner or executor behavior.
- Record the exact point where text becomes a tool argument, file path, network target, or secret request.
- Prove one minimal exploit chain before exploring variants.
- Keep prompt snippets and tool transitions in compact evidence blocks.
Workflow
1. Map The Control Stack
- Track system, developer, user, retrieved, memory, planner, and tool-response layers separately.
- Distinguish claimed capability from runtime-exposed capability.
- Note what the model can actually call, read, or mutate.
2. Prove The Boundary Crossing
- Reproduce one chain from untrusted text to changed planner behavior, changed tool args, or secret exposure.
- Keep the decisive transcript compact: source chunk, rewritten planner state, final tool invocation.
- Prefer the smallest transcript that still demonstrates the bug.
3. Report By Boundary
- State which layer failed: retrieval, summarizer, planner, executor, tool normalization, or output post-processing.
- Separate instruction drift from actual side effect.
Read This Reference
- Load
references/prompt-injection.mdfor the checklist, evidence layout, and common prompt-boundary pitfalls.
What To Preserve
- Original malicious chunk or prompt
- Intermediate summary or planner drift if it matters
- Final tool args, file paths, or exposed secret surface
04 同仓库更多 Skills
windows-ad
Use for authorized Active Directory and Windows identity attacks including Kerberos, AD CS, BloodHound paths, NTLM relay, and domain privilege escalation research.
wifi-wireless
Use for authorized wireless security assessment including Wi-Fi capture, WPA handshake analysis, rogue AP detection research, and lab-only deauth testing.
threat-hunting
Use for blue-team threat hunting, detection engineering with Sigma/YARA, SIEM query design, and incident detection validation.
thick-client
Use for authorized security testing of desktop thick clients including local storage, update channels, IPC, traffic, and client-side trust boundaries.
supply-chain-security
Use for software supply-chain security assessment covering SBOM, SCA, CI/CD pipelines, container images, build integrity, dependency provenance, and vulnerability reachability.
dsl-vm-reverse
Reverse JavaScript-based custom DSL/VM interpreters, non-standard WASM-like runtimes, and risk-control engines. Use when analyzing IIFE or switch-based opcode dispatchers, extracting instruction tables, recovering bytecode semantics, capturing VM state at runtime, or reconstructing execution flow.