The Agent Skills Directory

[DATA_EXPOSURE]: The skill instructions specify checking for placeholders and secrets within sensitive files, including .env, .env.local, and source code files. While these checks are intended for security verification, accessing these paths exposes highly sensitive credentials.
[INDIRECT_PROMPT_INJECTION]: The skill evaluates project files, such as qa-report.json, environment variables, and source code, as part of the deployment verification process. This ingestion of untrusted data, combined with the agent's ability to execute shell commands (Bash), creates a surface for indirect prompt injection. Ingestion points: .env, .env.local, qa-report.json, and project source code files (.py, .ts, .js). Boundary markers: No explicit delimiters or 'ignore' instructions are used when processing these files. Capability inventory: Full Bash shell execution, Read, Write, Glob, and Grep tools. Sanitization: The skill lacks explicit sanitization or validation of the content within the ingested project files.

deployment-checklist