The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by relying on instructions or data parsed from external files to guide its behavior.
Ingestion points: The skill reads specforge/state.md, tasks.md, and spec.md from the project directory to determine the current state and next action.
Boundary markers: None. The instructions do not define delimiters or provide the agent with warnings to disregard potential commands embedded within the project files.
Capability inventory: The skill utilizes Read, Glob, and Grep tools and is designed to prompt for user approvals or continue implementation tasks based on the file content.
Sanitization: Absent. There is no evidence of content validation, escaping, or filtering before the data is processed into the agent's context.

resume