The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions direct the agent to use npx @leoflores/datadog-cli, which downloads and executes code from the npm registry at runtime. This introduces a supply chain risk, as the package source is not from a verified or trusted vendor.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to search, stream, and aggregate Datadog logs. Since log content can be controlled by external users or services, this presents an indirect prompt injection surface where an attacker could log instructions to influence the agent's behavior.
Ingestion points: logs search, logs tail, logs patterns, and logs context commands (SKILL.md, references/logs-commands.md).
Boundary markers: Absent. The instructions do not provide delimiters or specific guidelines for the agent to ignore instructions embedded in the log data.
Capability inventory: The skill allows shell command execution via npx and file writing via the --output flag.
Sanitization: No sanitization or validation of log content is performed before it is presented to the agent.
[COMMAND_EXECUTION]: The dashboard update workflow in references/dashboards.md instructs the agent to extract JSON fields using jq into shell variables and then interpolate those variables into a new command line. This pattern is vulnerable to command injection if the data being processed contains shell metacharacters and the agent fails to escape them properly.

datadog-cli