The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The installation instructions in SKILL.md recommend the command curl -fsSL https://raw.githubusercontent.com/xdevplatform/xurl/main/install.sh | bash. This pattern downloads a remote script and pipes it directly into the shell for execution, which is an insecure practice that allows arbitrary code execution from a third-party source without user verification.
[EXTERNAL_DOWNLOADS]: The skill facilitates downloading code and binaries from untrusted sources, including the install.sh script from raw.githubusercontent.com/xdevplatform/xurl, the @xdevplatform/xurl package via NPM, and the github.com/xdevplatform/xurl repository via Go.
[INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data that could potentially contain malicious instructions intended to manipulate the agent's behavior.
Ingestion points: In SKILL.md, the commands xurl search, xurl timeline, xurl mentions, xurl read, and xurl dms ingest external content from the X (Twitter) API.
Boundary markers: Absent. The instructions do not define delimiters or provide specific prompts to ensure the agent treats the fetched data as non-executable text.
Capability inventory: In SKILL.md, the skill includes high-privilege write capabilities such as xurl post, xurl reply, xurl dm, xurl follow, xurl block, and raw API access through xurl -X.
Sanitization: Absent. The skill returns raw JSON data from the API to the agent context without applying filtering, escaping, or validation logic to sanitize potentially malicious payloads.

xurl