This skill is a specialization of the ctf-sandbox-orchestrator skill for challenges where runtime truth lives in built assets. The agent starts from the served artifact set, which means the entry HTML, build manifest, bootstrap bundle, chunk map and source maps, and reconstructs the smallest runtime graph that explains what executes.
It keeps the served artifacts apart from repository source until the two are shown to match. The skill states that it runs only after the orchestrator has set sandbox assumptions.
What You Get
- A three-step workflow: map the served artifact set, reconstruct runtime structure, reduce to the decisive bundle path
- A reference checklist covering entry HTML, manifests, chunk registries, source map URLs and framework route manifests
- A rule to check manifests and source maps before opening large minified bundles
- A hand-off rule to the template-render skill when the task moves to server-side rendering
Common Use Cases
- Recovering hidden routes or API endpoints from the emitted assets of a CTF web app
- Explaining which chunk a lazy loader pulls and what it contains
- Checking that the repository source matches the bundle the server delivers
What Makes It Unique
It treats the served artifacts as the source of truth and asks the agent to prove which bundle path executes. Its checklist warns against trusting repository source over the served asset set, and against recovering names without showing which module runs.
Under The Hood
The skill is instruction text only: a SKILL.md, one reference checklist and an agent configuration file. It ships no scripts. The method involves requesting the served pages and assets, reading source maps and manifests, and running or decoding bundle code, so the agent’s own tools fetch pages and run commands in the sandbox the orchestrator set up. The skill itself does not call APIs or write files.