The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local script or binary at $HOME/.dotfiles/bin/spatie-ops to perform authenticated API calls.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes DNS zone records from the Openprovider REST API, which presents a potential surface for injection if malicious instructions are embedded in the record values.
Ingestion points: The skill performs a GET /dns/zones/$ZONE?with_records=true call in SKILL.md to retrieve external data.
Boundary markers: There are no specific delimiters or instructions provided to the agent to disregard potential instructions embedded within the DNS record data.
Capability inventory: The skill has capabilities to write files (zone backups) and execute shell commands via the spatie-ops helper and jq.
Sanitization: The skill uses jq to extract specific fields like type and name, providing structural parsing but no sanitization of the values themselves.

openprovider-dns