The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/reload-headscale-policy.sh executes several shell commands, including kill, pgrep, ps, and journalctl, to manage the Headscale service process and verify reload status.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external huJSON policy files, which presents a surface for indirect instructions to influence agent behavior. * Ingestion points: User-provided policy files are read by the agent and processed by validate-policy.py and migrate-acls-to-grants.py. * Boundary markers: The skill instructions do not specify the use of delimiters (e.g., XML tags) or provide explicit warnings for the agent to ignore instructions embedded within the policy data. * Capability inventory: The skill is capable of reading and writing local files and executing shell commands via the included automation scripts. * Sanitization: The skill lacks mechanisms to sanitize or filter the content of processed files to prevent the execution of malicious natural language instructions.

tailnet-policy