The Agent Skills Directory

[COMMAND_EXECUTION]: The documentation in SKILL.md includes multiple functional shell command examples using curl and grep to analyze HTTP headers and test for path-based cache deception.
[DATA_EXFILTRATION]: The scripts/agent.py script performs outbound network requests to a user-provided target_url using the requests library. This is intended for the primary purpose of identifying caching vulnerabilities in web applications.
[PROMPT_INJECTION]: This skill presents an indirect prompt injection surface (Category 8).
Ingestion points: The scripts/agent.py script ingests the full response body of external web pages when testing for path confusion and PII exposure.
Boundary markers: The skill does not implement boundary markers or instructions to isolate the ingested external content from the agent's processing logic.
Capability inventory: The script uses requests.get to fetch external content and print to generate JSON reports containing data from those responses.
Sanitization: No sanitization or escaping is performed on the data fetched from external URLs before it is output, which could allow malicious content from a target site to influence the agent's behavior.

performing-web-cache-deception-attack