The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, potentially untrusted sources such as local error logs and production error monitoring systems (Sentinel).
Ingestion points: Processes data from local error logs, deployment subscriptions, and Sentinel prod-error rows as described in the Workflow section of SKILL.md.
Boundary markers: The skill instructions do not specify any delimiters or safety markers to differentiate between legitimate event data and potential malicious instructions embedded within that data.
Capability inventory: The skill workflow explicitly directs the agent to "fix" errors or "build" features based on the event data, implying capabilities for file modification and command execution.
Sanitization: No sanitization or validation logic is mentioned to ensure that event descriptions or error logs do not contain prompt injection payloads.

convex-monitor