The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill utilizes npx @leoflores/datadog-cli for its core operations. This pattern downloads and executes a package from the NPM registry at runtime. The package belongs to a third-party individual rather than a trusted vendor or official organization, creating a dependency on unverifiable code.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display Datadog logs, which can contain untrusted data originating from external users or systems. Without proper boundary markers or sanitization, malicious instructions embedded in these logs could be processed as commands by the agent. -- Ingestion points: The logs search and logs tail commands described in references/logs-commands.md and SKILL.md pull log data into the agent context. -- Boundary markers: No explicit delimiters or instructions are provided to the agent to treat log data as untrusted text. -- Capability inventory: The skill uses npx for CLI operations and jq for JSON processing, allowing for shell-based data manipulation. -- Sanitization: There is no evidence of filtering or escaping logic applied to the log content before it is read by the agent.
[COMMAND_EXECUTION]: Recommended workflows in references/dashboards.md involve shell command substitution and local file operations (e.g., TEMPLATE_VARS=$(jq ... /tmp/dashboard.json)). This encourages the agent to perform complex shell manipulations on data fetched from a remote API, which could lead to command injection if the API responses are maliciously crafted.
[EXTERNAL_DOWNLOADS]: The skill instructions frequently trigger downloads of the @leoflores/datadog-cli package via npx. This creates a recurring dependency on the availability and integrity of an external third-party repository.

datadog-cli