The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to interact with the roam-cli utility and perform system operations such as binary installation.\n- [EXTERNAL_DOWNLOADS]: The skill's installation guide in references/installation.md directs users to download the roam-cli binary from the author's GitHub repository.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its ingestion of data from external Roam Research pages.\n
Ingestion points: Commands in SKILL.md like get, search, and journal read untrusted content from the Roam graph.\n
Boundary markers: There are no markers or instructions to isolate retrieved content from the agent's context.\n
Capability inventory: The skill allows significant modifications to the graph via save, block create, and batch run as specified in SKILL.md.\n
Sanitization: No evidence of data sanitization or validation is present in the skill's logic.

roamresearch