The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, which creates a potential surface for indirect prompt injection attacks.
Ingestion points: Instructions in SKILL.md (Step 1) and resources/template.md direct the agent to gather facts from potentially untrusted sources like system logs, metrics, and witness accounts.
Boundary markers: The skill lacks explicit instructions or delimiters to help the agent distinguish between data to be analyzed and instructions to be followed within those data sources.
Capability inventory: The skill's potential for harm is strictly limited as it contains no capabilities for network operations, arbitrary shell command execution, or writing to sensitive file system locations.
Sanitization: There are no mentioned mechanisms for sanitizing, validating, or filtering the content of the external logs or reports before they are processed by the agent.

postmortem