The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the allowed-tools configuration to implement the principle of least privilege, restricting the agent to a specific whitelist of pup CLI subcommands rather than allowing arbitrary shell access.
[CREDENTIALS_UNSAFE]: While the skill requires sensitive environment variables (DD_API_KEY, DD_APP_KEY), it contains explicit instructions to the agent to verify their existence without ever revealing or printing their actual values.
[EXTERNAL_DOWNLOADS]: The skill references the official GitHub repository for the pup CLI (github.com/datadog-labs/pup). As Datadog is a well-known and established technology service provider, this reference is documented neutrally.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data that could be influenced by third parties (such as log messages or monitor names).
Ingestion points: Data enters the context via pup logs search, pup monitors list, and other query commands in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or boundary markers for external data processing.
Capability inventory: The skill has the capability to write to the local filesystem (Write tool) and execute a defined list of pup commands.
Sanitization: There is no explicit sanitization step described for the retrieved observability data before it is processed for the report.

datadog-analyser