The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external metadata, such as update messages and branch names, retrieved from the EAS service via eas-cli. This data could be manipulated by collaborators to include instructions, creating a surface for indirect prompt injection.
Ingestion points: Data returned by eas update:list and eas channel:insights as documented in SKILL.md.
Boundary markers: The skill does not provide explicit delimiters or instructions for the agent to ignore or isolate potential instructions embedded within the update metadata.
Capability inventory: The skill allows the execution of eas commands via the Bash tool, as restricted by the allowed-tools frontmatter.
Sanitization: No explicit sanitization or filtering of the retrieved metadata is mentioned before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill includes instructions for a feedback command using npx --yes @expo/agent-cli@latest. This command downloads and executes a package from the author's official registry.
Evidence: The command is found in the 'Submitting Feedback' section of SKILL.md.

eas-update-insights