The Agent Skills Directory

[PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill is designed to process user-provided data to construct database queries, creating a potential vector for indirect prompt injection if the data contains malicious instructions.\n- Ingestion points: Data enters through user-provided variables such as workspaceId, since, and name used in Prisma Client and raw SQL examples in SKILL.md.\n- Boundary markers: Absent. The skill does not define specific markers (e.g., XML tags or delimiters) to isolate untrusted data within the prompts it generates.\n- Capability inventory: The skill possesses the capability to perform database reads and writes using Prisma ORM and $queryRaw methods.\n- Sanitization: Present. The skill mandates the use of parameterized queries and explicitly warns against building SQL strings from untrusted input, which significantly mitigates direct SQL injection risks.\n- [COMMAND_EXECUTION] (SAFE): No evidence of system-level command execution or shell script spawning.\n- [DATA_EXFILTRATION] (SAFE): No patterns of network operations or sensitive file access (e.g., SSH keys, credentials) were found.\n- [CREDENTIALS_UNSAFE] (SAFE): No hardcoded API keys, tokens, or database connection strings are present.\n- [REMOTE_CODE_EXECUTION] (SAFE): The skill does not download or execute remote scripts; it uses standard library patterns for database interaction.

prisma-database-querying