The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of the glean CLI tool to perform document operations. The commands and flags used are consistent with the skill's stated purpose of document retrieval and inspection.
[PROMPT_INJECTION]: The skill ingests and processes document content and metadata from an external indexing service (Glean). This creates a surface for indirect prompt injection where malicious instructions hidden in indexed documents could attempt to influence the agent. This is inherent to the primary purpose of the skill. * Ingestion points: SKILL.md (subcommands get and summarize process document data). * Boundary markers: Not specified in this file. * Capability inventory: The skill uses a specialized CLI tool but does not demonstrate dangerous execution capabilities like eval or sudo. * Sanitization: No specific document content sanitization is described.

glean-documents