The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because its core logic involves reading and interpreting the contents of numerous project files which could be manipulated to influence the agent's routing decisions.
Ingestion points: The skill reads file content from research/, brand/, architecture/, .agents/, and CLAUDE.md to build a project state map (identified in SKILL.md Step 1).
Boundary markers: There are no explicit instructions to use delimiters or security-focused parsing logic to ignore instructions that might be embedded within the processed project files.
Capability inventory: The skill utilizes Read, Grep, Glob, and Bash tools. Its primary capability is recommending further commands to the user, such as /start-research or /fresh-eyes.
Sanitization: The instructions do not mention any sanitization, filtering, or validation steps for the data retrieved from the local repository before it is used to determine the workflow path.

start-meta