The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface.
Ingestion points: The skill retrieves untrusted data from PagerDuty incident lists, alert details, and notes via tools like PAGERDUTY_FETCH_INCIDENT_LIST and PAGERDUTY_GET_ALERTS_BY_INCIDENT_ID described in SKILL.md.
Boundary markers: No delimiters or safety instructions are provided to isolate external PagerDuty data from the agent's core instructions.
Capability inventory: The skill enables modification of PagerDuty resources, such as escalation policies and on-call schedules, which could be abused if malicious instructions are processed.
Sanitization: No sanitization or validation of PagerDuty API responses is performed before the data is processed or used in subsequent actions.
[DATA_EXFILTRATION]: External Network Dependency.
The skill instructs users to configure an external MCP server at https://rube.app/mcp. This setup results in the transmission of agent context and interaction data to an external service provider that is not included in standard whitelists.

pagerduty-automation