The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's installation instructions include a command that downloads a shell script from a remote GitHub repository and pipes it directly into the bash interpreter. This pattern allows for arbitrary code execution from an untrusted source.
Evidence: curl -fsSL https://raw.githubusercontent.com/geekjourneyx/jina-cli/main/scripts/install.sh | bash in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill prompts the user or agent to install software from an external repository and interacts with external APIs to fetch content.
Evidence: go install github.com/geekjourneyx/jina-cli/cli@latest and references to https://r.jina.ai/ and https://s.jina.ai/ in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill provides functionality to fetch content from arbitrary URLs and search results, which is then processed into a format intended for LLM consumption. This creates a surface where external content can influence agent behavior.
Ingestion points: The jina read --url, jina search --query, and jina read --file commands documented in SKILL.md.
Boundary markers: No delimiters or specific instructions to ignore embedded commands are present in the skill text.
Capability inventory: The skill involves executing CLI commands that perform network operations and local file reading.
Sanitization: The tool converts raw content to Markdown, but there is no evidence of sanitization to prevent instruction injection within the data.
[PRIVILEGE_ESCALATION]: The skill provides instructions to modify the system's environment variables during the Windows setup process.
Evidence: [Environment]::SetEnvironmentVariable("Path", $env:Path + ";$HOME\go\bin", "User") in SKILL.md.
[COMMAND_EXECUTION]: The skill contains multiple shell command examples intended to be executed by the agent to install and use the CLI tool.

jina-cli