The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes alerts and templates from external sources, creating an indirect prompt injection surface where malicious content in an alert description could attempt to influence the agent's behavior.
Ingestion points: Alert event data from monitoring platforms (e.g., Datadog) and template page content from wiki platforms (Confluence/Notion) are read into the context in SKILL.md.
Boundary markers: Absent. The instructions do not specify the use of delimiters or 'ignore' instructions for the external content.
Capability inventory: The skill has access to the Bash tool and wiki write operations.
Sanitization: Absent. There are no instructions to escape or validate data from monitoring alerts before it is written to the wiki.
[COMMAND_EXECUTION]: The skill's configuration requests the Bash tool. While no specific shell commands are provided in the workflow, the availability of a shell environment alongside untrusted input from external alerts increases the potential risk profile.
[DATA_EXFILTRATION]: The skill's primary function is to read sensitive operational data (alerts) from monitoring tools and write it to a wiki. Users should ensure that the destination wiki is a trusted internal resource to prevent exposure of system health data.

ops-oncall-log