The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted project files such as configuration manifests and dependency definitions, creating a surface for indirect prompt injection.
Ingestion points: Medication management platform configurations, dependency manifests, and environment definitions are read from the target project's file system.
Boundary markers: The skill lacks delimiters or instructions to prevent the agent from executing malicious prompts embedded within the files being analyzed.
Capability inventory: The skill utilizes file system read operations across the project and write operations for telemetry.
Sanitization: No sanitization or validation of the content within the analyzed files is performed before processing.
[DATA_EXFILTRATION]: The skill explicitly directs the agent to read 'environment definitions' and 'configuration files.' These locations are standard targets for sensitive information such as API keys, database credentials, and secrets. It also targets highly sensitive Protected Health Information (PHI), including medication records, adverse event reports, and patient consent data.
[PERSISTENCE]: The 'Self-Evolution Telemetry' section instructs the agent to write execution metadata to a file located at ~/.claude/projects/skill-telemetry.md. Appending data to a hidden directory in the user's home folder constitutes a persistence mechanism that tracks activity and logs state across different sessions and projects.

medication-adherence