The Agent Skills Directory

Data Exposure (HIGH): The SKILL.md file reveals the full path to a sensitive .env file containing API keys in the user's iCloud directory (~/Library/Mobile Documents/com~apple~CloudDocs/Geoffrey/secrets/.env). This disclosure assists attackers in locating sensitive credentials on the local filesystem.
Indirect Prompt Injection (LOW): The skill fetches and summarizes untrusted ticket data (subjects and descriptions) from Freshservice, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: Ticket subjects and descriptions retrieved via API in get_daily_summary.js, get_weekly_summary.js, and get_ticket.js.
Boundary markers: Absent. Ticket content is concatenated and processed without delimiters or 'ignore' instructions.
Capability inventory: The skill has tools to create tickets (create_ticket.js), update tickets (update_ticket.js), and add notes (add_note.js).
Sanitization: Absent. No filtering or escaping is applied to the ticket data before it is used in logic or summaries.
Unverifiable Dependencies (MEDIUM): All script files (scripts/*.js) import secrets from a hardcoded relative path ../../../scripts/secrets.js. This dependency resides outside the skill's directory structure, creating a non-standard and potentially risky dependency on files in the parent environment's directory tree.

freshservice-manager