The Agent Skills Directory

Indirect Prompt Injection (INFO): The skill is designed to ingest and process untrusted data from changelogs and external web sources, creating a potential surface for indirect prompt injection.
Ingestion points: Processes the changelogs array from input and fetches content from external URLs via WebSearch and WebFetch.
Boundary markers: The instructions do not specify delimiters or boundary markers to isolate external content from the agent's instructions.
Capability inventory: The skill is limited to information retrieval and generating a user-friendly summary for display. It does not possess file-write, command execution, or data exfiltration capabilities.
Sanitization: No sanitization or content filtering is implemented for the retrieved data.
External Network Access (LOW): The skill performs network operations to gather feature details.
Source status: Targets are restricted to trusted official sources (anthropic.com and github.com), which downgrades the severity of the network operations per the [TRUST-SCOPE-RULE].

changelog-interpreter