How to use the Competition Template Render Path agent skill

What does the skill make the agent do?

The skill gives the agent a structured method for analysing server-side rendering (SSR) and template routes in a sandboxed competition target. The agent follows a render chain in order: route resolution, loader or data fetch, template or component context, response HTML, hydration payload and client takeover.

It also checks whether template-level checks such as permissions, feature flags, preview state or host-based switches match the backend's own enforcement. The source says the agent should not trust any one layer on its own, and should record hidden fields, inline data, hydration JSON, meta tags or alternate partials that expose the decisive branch.

• Maps the render chain for a view: route, loader, template, HTML, hydration data and client boot.
• Compares template gating with loader and handler enforcement.
• Reduces the finding to one minimal request-to-render sequence.
• Keeps one successful render and one failing render path to show where behaviour diverges.

Before you install: the parent skill is required

This skill is a downstream specialisation. The source says it should only be used after the parent skill ctf-sandbox-orchestrator is already active and has established sandbox assumptions, node ownership and evidence priorities. If that has not happened, the agent is told to return to the parent first.

It is not a standalone tool. Install and activate ctf-sandbox-orchestrator before you rely on this skill.

How to install the skill

Download the ZIP file from this page. Then choose one of the two options below, depending on how you use your agent.

Option one is a personal install: unzip the file into ~/.claude/skills/. Option two is a project install: unzip it into the .claude/skills/ folder of a specific project. You can also upload the ZIP in the Claude apps under Settings, then Capabilities, then Skills.

• Personal: unzip into ~/.claude/skills/
• Project: unzip into .claude/skills/ inside your project
• Claude apps: Settings, then Capabilities, then Skills, then upload the ZIP

How to trigger the skill with natural requests

You do not need a special command. The skill's description lists the situations where it should activate. Ask for one of those tasks in plain language, and make sure the sandbox context from the parent skill is already in place.

The source lists these trigger situations: inspecting SSR or template routes, tracing render context or hydration data, comparing template gating with handler enforcement, explaining preview or hidden-route rendering, and connecting render pipeline behaviour to the decisive branch.

• Inspect the SSR routes of this sandbox target and tell me which loader feeds the template.
• Trace the hydration data on the product page and check whether any hidden fields are exposed.
• Compare the template's admin gating with what the backend handler enforces.
• Explain how the preview route renders hidden content for this challenge.

What the skill asks you to provide and what it produces

The skill needs a sandboxed competition target to analyse. The source does not specify particular tools or accounts, so it does not name any you must have. Keep the work inside the authorised sandbox.

The agent is told to record host, route params, preview toggles, tenant or host switches and server-only variables before changing anything. It packages evidence as route-side, render-side and effect-side notes, so the result can be followed step by step.

• Route side: host, path, matched route, params, loader or handler.
• Render side: template or layout, server-only variables, hydration keys, hidden fields.
• Effect side: rendered content, leaked data, privileged action or bypassed branch.

Safety notes: scope, scripts and network

The skill is written for sandboxed competition targets. It is not meant for live production systems or third-party sites without authorisation. Check that your target falls within that scope before you start.

According to the listing, this skill runs no scripts and uses no network access. Bundle recovery and source map reconstruction are outside its scope. The skill says to hand those tasks to a different, tighter bundle skill.

• Use only against sandboxed competition targets you are authorised to test.
• Do not use it on live production systems or third-party sites without permission.
• Bundle recovery or source map reconstruction should go to a different skill.

Tips for better results

Start by establishing sandbox assumptions with the parent skill. Then ask the agent for a narrow question, such as one route or one template, rather than the whole application. The checklist in the skill's reference file asks the agent to look at hydration JSON and inline data blobs as well as the visible HTML.

The skill also asks for a language setting. By default it replies in Simplified Chinese, unless you explicitly ask for English in your request.

• Ask about one route or one template at a time.
• Mention the specific hidden field, hydration key or loader you care about.
• Ask for both a successful and a failing render path to compare.
• Request English in your message if you do not want Simplified Chinese replies.