This skill is a specialization of the ctf-sandbox-orchestrator skill for challenges where the decisive edge is proving how metadata-derived identity becomes accepted privilege. The agent identifies the active metadata surface, records the reachability path, and separates three claims: metadata is reachable, a credential was issued, and a downstream API accepted it.
The skill states that it runs only after the orchestrator has set sandbox assumptions.
What You Get
- A three-step workflow: map metadata reachability, prove credential or identity issuance, reduce to the decisive trust path
- A reference checklist for reachability, token claims and the accepting API
- An evidence list covering endpoints, headers, role names, audiences, issuers and privilege-bearing actions
- Hand-off rules to the SSRF pivot skill and to the Kubernetes control-plane skill
Common Use Cases
- Documenting how a pod or instance in a lab obtains a token from its metadata service
- Explaining which role or audience a metadata-derived credential carries
- Writing up the chain from a reaching surface to an accepted cloud action in a CTF
What Makes It Unique
It treats reaching the metadata service as only the first link and requires proof of issuance and acceptance. Its checklist warns about two gaps: showing metadata access without a useful credential, and mixing node identity with workload identity without showing which one drove the privilege edge.
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 calling metadata endpoints and cloud or cluster APIs from within the sandbox, so the agent’s own shell and network tools do that work under the assumptions the orchestrator set up. The skill itself does not fetch pages, call APIs or write files.