The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous curl command examples in the detailed guide (references/detailed-guide.md) for sending HTML injection payloads to target servers. These are intended for security testing and are subject to the mandatory confirmation gate defined in the skill's root file.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest and process data from external web applications (responses from curl and Python requests).
Ingestion points: The skill reads response text from user-specified URLs in the Python fuzzing script and curl commands (references/detailed-guide.md).
Boundary markers: The skill includes a "Mandatory confirmation gate" in SKILL.md requiring explicit user authorization and scope confirmation before execution.
Capability inventory: The skill has the capability to execute shell commands (curl) and run Python code that performs network requests (requests.get).
Sanitization: The provided Python script performs simple string matching on the response body to detect reflected payloads rather than executing or interpreting the response as instructions.

html-injection-testing