The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted data from external Roam Research pages which may contain malicious instructions.\n
Ingestion points: Content is retrieved via roam-cli get, search, and journal commands as documented in SKILL.md and usage-examples.md.\n
Boundary markers: There are no instructions for using delimiters or boundary markers to isolate processed data from the agent's instructions.\n
Capability inventory: The skill provides write and modification capabilities through roam-cli save, block create, and batch run as defined in SKILL.md.\n
Sanitization: No sanitization or validation routines are specified for handling retrieved content before it is processed by the agent.\n- [EXTERNAL_DOWNLOADS]: The skill requires downloading and installing a CLI tool from a remote repository.\n
Download source: Installation instructions in references/installation.md direct the user to download roam-cli from the Leechael/roamresearch-skills repository on GitHub.\n
Installation process: Requires extracting a tarball and installing the binary into a system path (/usr/local/bin).\n- [CREDENTIALS_UNSAFE]: The skill manages sensitive authentication data through environment variables.\n
Sensitive keys: Relies on ROAM_API_TOKEN and ROAM_API_GRAPH for API access.\n
Access method: Credentials are provided via environment variables or injected using tools like 1Password CLI as described in references/installation.md.\n- [COMMAND_EXECUTION]: Core functionality is delivered through numerous subprocess executions of the roam-cli binary with various arguments and piped inputs.

roamresearch