The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to guide the agent in executing the glean CLI tool and jq for searching, reading, and managing enterprise documents.
[EXTERNAL_DOWNLOADS]: The documentation suggests installing the glean CLI using brew install gleanwork/tap/glean-cli. This reference points to the official distribution source for the Glean technology vendor.
[DATA_EXFILTRATION]: The skill involves handling sensitive corporate data and authentication tokens. It includes a specific security instruction to avoid logging or echoing the GLEAN_API_TOKEN and recommends secure storage methods like the system keyring.
[PROMPT_INJECTION]: The skill facilitates the ingestion of content from various corporate sources, which presents a surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through commands like glean search, glean chat, and glean documents as described in SKILL.md and references/commands.md.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore instructions embedded within the retrieved enterprise content.
Capability inventory: The agent is granted the ability to execute a wide range of subcommands via glean, including write operations such as announcements create and collections update.
Sanitization: There are no instructions for sanitizing or filtering the content fetched from external document sources before it is processed by the agent.

glean-cli