The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process untrusted data from external sources, which creates a vulnerability surface for indirect prompt injection.
Ingestion points: The 'Voice calibration' phase in SKILL.md instructs the agent to use WebFetch to read content from user-supplied Substack URLs for tone and style analysis.
Boundary markers: The instructions do not specify any delimiters or boundary markers to isolate the fetched external content from the agent's instructions, nor do they include 'ignore' directives for embedded commands.
Capability inventory: The skill allows the agent to use Read, Write, Edit, WebFetch, and Agent (cross-agent invocation) tools, providing a significant capability set that could be targeted by injected instructions.
Sanitization: There is no evidence of sanitization, filtering, or validation performed on the retrieved content before it is processed by the agent.

substack-ghostwriting