The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to download and execute code from a third-party NPM package @leoflores/datadog-cli using npx. This package is not an official release from the service provider.
[COMMAND_EXECUTION]: All skill capabilities are implemented via shell commands. The workflow for updating dashboards in references/dashboards.md involves capturing JSON data into shell variables and then interpolating those variables back into new shell commands, which creates a risk of command injection if the data contains malicious shell characters.
[DYNAMIC_EXECUTION]: The skill uses jq to dynamically parse and manipulate JSON data at runtime, then passes the resulting strings as arguments to shell commands. This pattern of dynamic command construction is susceptible to manipulation by malformed or malicious input data.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display untrusted data from external Datadog logs and metrics, which could potentially contain adversarial instructions for the agent.
Ingestion points: logs search, logs tail, logs trace, logs context, and logs patterns commands (found in SKILL.md and references/logs-commands.md).
Boundary markers: None present; there are no instructions to the agent to ignore or delimit embedded text in log messages.
Capability inventory: Shell command execution via npx, environment variable access for API keys, and file system writes via the --output flag.
Sanitization: No evidence of data sanitization or escaping of log/metric content before display or reuse in commands.

datadog-cli