The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions include commands to download and execute scripts directly from a remote server to install the required CLI tool.
Evidence: curl -fsSL https://d.officecli.ai/install.sh | bash for macOS/Linux and irm https://d.officecli.ai/install.ps1 | iex for Windows.
The resources are hosted on d.officecli.ai, which is the official domain for the tool specified in the skill's author context.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external CSV and tabular data, which creates a potential surface for indirect prompt injection attacks.
Ingestion points: sales.csv imported using the officecli import command in Phase 1.
Boundary markers: The skill does not explicitly use delimiters or instruction-bypass warnings for the imported data content.
Capability inventory: The skill performs file creation, metadata modification (using raw-set with XML templates), data visualization (charts, sparklines), and shell command execution via the officecli binary.
Sanitization: The skill documentation mentions inheriting shell quoting and execution discipline from officecli-xlsx to mitigate command injection risks.
[COMMAND_EXECUTION]: The skill relies extensively on shell command execution to interact with the file system and the officecli tool, including complex piping and command substitution.
Evidence: Multiple bash blocks using officecli, jq, and shell variables (e.g., DASH_IDX=$(officecli query ... | jq ...)).

officecli-data-dashboard