The Agent Skills Directory

[SAFE]: The skill's code and instructions are consistent with its stated purpose as a CTF web analysis tool. No malicious patterns targeting the user's environment or credentials were detected. The exploitation code provided in references is for educational and competitive use within the context of a CTF challenge.
[COMMAND_EXECUTION]: The skill provides reference materials and instructions for using command-line tools like curl to interact with target web applications for vulnerability verification and state analysis (e.g., in references/cookie-hmac-key-reuse-auth-bypass.md).
[PROMPT_INJECTION]: The skill has an inherent attack surface for indirect prompt injection because it is designed to ingest and analyze untrusted external data, such as HTML, scripts, and browser storage content from target sites (specified in SKILL.md and references/routing-runtime.md).
Ingestion points: External web responses, boot scripts, cookies, and localStorage data analyzed during CTF tasks.
Boundary markers: Not explicitly defined for separating untrusted data from instructions.
Capability inventory: Shell command execution (curl) and system inspection tools.
Sanitization: The skill does not provide explicit sanitization logic for the data it processes, relying on the user's sandbox orchestrator for safety.

competition-web-runtime