The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from client engagement logs and meeting notes.
Ingestion points: In SKILL.md, the agent is instructed to discover and read external files including CSV, JSON, CRM exports, meeting notes, and communication logs.
Boundary markers: The instructions do not define boundary markers (such as XML tags or delimiters) or provide explicit directives to ignore instructions that might be embedded within the ingested text.
Capability inventory: The skill possesses the Write tool to create the renewal-forecast.md file and has Bash tool permissions for discovery operations.
Sanitization: There are no instructions provided for filtering, escaping, or sanitizing text from external logs before it is used for health scoring and risk assessment.

renewal-predictor