Guide · Agent skill
How to use the Competition Relay Coercion Chain agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
Competition Relay Coercion Chain is an add-on skill for an AI agent working on capture-the-flag (CTF) security competitions in an isolated lab. It helps the agent turn a forced-authentication step into a traced chain: what triggers the authentication, where it is relayed, which service accepts it, and what privilege or artifact results. This guide explains what the skill does, how to install it, how to trigger it, and what it needs before it runs.
Competition Relay Coercion Chainzhaoxuya520/reverse-skill · MIT Download the skill Quick start
- 1Confirm the parent skill. Install and activate ctf-sandbox-orchestrator first, since this skill only runs after it has set the sandbox assumptions.
- 2Download the ZIP. Download the ZIP of the competition-relay-coercion-chain folder from this page.
- 3Unzip into a skills folder. Place the folder in ~/.claude/skills/ for personal use, or in .claude/skills/ inside a project.
- 4Or upload in the Claude apps. Go to Settings → Capabilities → Skills and upload the skill there.
- 5Start a lab session with a clear request. Describe the coercion primitive, relay path, or accepting service you want traced, inside your authorized sandbox.
What the skill makes the agent do
The skill organizes an attack-style analysis into one chain with five parts: the coercion source, the captured authentication, the relay target, the acceptance point, and the resulting effect. For each hop it asks the agent to record the transport, protocol, and service identity.
It is a downstream specialist. It is written to run only after the parent skill ctf-sandbox-orchestrator has set up the sandbox assumptions and routed the task here. By default the agent replies in Simplified Chinese unless you explicitly ask for English.
- Maps the coercion source and the trigger that makes a service authenticate
- Traces where the authentication lands and which service accepts the relayed identity
- Keeps trigger, relay transcript, and accepting service in one evidence chain
- Reduces the result to the smallest sequence that proves the decisive step
What you need before you install
This skill depends on another skill. The parent skill ctf-sandbox-orchestrator must be installed and active first. If it has not established sandbox assumptions, the agent is instructed to return to that parent skill before using this one.
The folder includes the files agents/openai.yaml, references/relay-coercion-chain.md, and SKILL.md. The reference file must stay in the folder, because the skill loads it for the coercion checklist, relay checklist, and evidence packaging. The skill does not run scripts and does not use the network.
- Parent skill: ctf-sandbox-orchestrator, installed and active
- Reference file: references/relay-coercion-chain.md, included in the folder
- No bundled scripts and no network use
Install the skill from the ZIP
Download the ZIP of the competition-relay-coercion-chain folder from this page. Unzip it so the competition-relay-coercion-chain folder sits inside a skills directory. For a personal install, use ~/.claude/skills/. For a single project, use the .claude/skills/ folder in that project.
If you use the Claude apps instead, upload the skill under Settings → Capabilities → Skills. Install the parent skill ctf-sandbox-orchestrator the same way so both are available.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside your project
- Claude apps: Settings → Capabilities → Skills, then upload
Trigger the skill with natural requests
You do not need a special command. The skill is meant to be picked up when your request matches its description. Phrasing that fits includes asking the agent to trace a coercion primitive, follow a relay path, analyze forced authentication, determine which service accepts relayed auth, or connect a coercion step to resulting privilege, enrollment, or code execution.
Start the session with the parent skill's sandbox context first. This skill is written to be used after that context exists, so a request made before the sandbox assumptions are set may be routed back to the parent.
- Trace this coercion primitive from the trigger to the source host.
- Follow the relay path and tell me which service accepts the relayed auth.
- Analyze this forced authentication event and list the relay target and protocol.
- Connect this coercion step to the resulting privilege in our lab.
Work through the chain step by step
The skill follows three workflow stages. First it maps the coercion source and records the host, coerced principal, transport, and preconditions. Next it traces the relay target, separating capture-only and replay-only results from actual relay acceptance. Finally it reduces everything to the smallest chain: coercion trigger, relayed auth, accepted service, and resulting privilege or artifact.
The reference checklist warns against common errors: stopping at 'forced auth happened' without proving relay acceptance, proving acceptance without showing what it produced, and mixing several candidate targets without isolating the one that accepted the auth.
- Keep the transcript, error or acceptance response, and exact protocol for each hop
- State whether the decisive weakness is in the coercion source, relay target, signing settings, or downstream service
- If the path becomes a certificate-enrollment issue or a pure Kerberos delegation edge, the skill says to hand off to a more specialized skill
Safety notes: scope and network
This skill is for isolated CTF labs and reproducible training exercises. Do not use it against real networks, systems, or accounts you do not own or lack written authorization to test. The skill traces relay and coercion paths that can cause real harm outside a sandbox.
The skill itself runs no scripts and makes no network calls. Any effect on a network comes from the commands and tools you or the agent run in your environment, so keep those inside the lab boundary the parent skill defines.
- Only use it in a sandbox you control or are authorized to test
- It focuses on exploitation paths, not defensive hardening
- Runs scripts: no. Uses network: no
Tips for better results
Give the agent a clear target list and sandbox context from the start, because the skill expects those assumptions to exist. Ask for evidence to be kept together per hop so the final write-up is easy to replay.
If you are a trainer, use the smallest coercion-to-acceptance sequence as the exercise. It is easier to check and reproduce than a full multi-target chain.
- Name the parent sandbox setup in your first message
- Ask for the smallest replayable sequence
- Request the reply in English if you do not want Simplified Chinese
What people use it for
- Tracing how a coercion trigger leads to a relayed authentication in a practice CTF lab
- Identifying which service in a lab accepts relayed auth and recording the transcript
- Building a reproducible training exercise on forced authentication and relay concepts for students
- Packaging evidence per hop so a CTF write-up can be checked step by step
Frequently asked questions
Is the Competition Relay Coercion Chain skill free?
The skill is published under the MIT licence, as stated on this page. Check the licence file in the downloaded folder for the full terms.
What do I need before installing it?
You need the parent skill ctf-sandbox-orchestrator installed and active, and the reference file references/relay-coercion-chain.md must stay in the folder. The skill runs no scripts and uses no network.
Where do I install agent skills?
For personal use, place the folder in ~/.claude/skills/. For one project, use .claude/skills/ in that project. In the Claude apps, upload it under Settings → Capabilities → Skills.
Why doesn't the skill respond in English?
The skill tells the agent to reply in Simplified Chinese unless the user explicitly requests English. Ask for English in your request if you want it.
Can I use this skill on a real network or a live system?
No. It is built for isolated CTF sandboxes and training labs. Do not use it on systems or accounts you do not own or have written authorization to test.
How is this different from a defensive security checklist?
It focuses on exploitation paths, tracing how coercion, relay, and acceptance connect. It is not a general hardening or defence guide, so use a defensive resource for that.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
This skill suits security students and CTF players, and red-team trainers building reproducible lab exercises on forced authentication and relay, provided they have the parent skill and an isolated sandbox. It is a poor fit for anyone testing real networks without written authorization, or for anyone who wants a defensive checklist.
Download the skill