The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection due to its handling of untrusted user input combined with high-privilege tool access.\n- Ingestion points: Untrusted data is ingested via the $ARGUMENTS variable in SKILL.md.\n- Boundary markers: Absent; the instructions do not define delimiters or specific 'ignore' rules for content within the user request.\n- Capability inventory: The agent is granted Write and Bash tool access, which allows for file system changes and shell command execution (referencing the allowed-tools in SKILL.md frontmatter).\n- Sanitization: Absent; no methods for sanitizing or validating the user-provided input are specified before the data influences agent actions.

moonshine-field