The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes shell commands at load time (e.g., ls, git log, go list) to inject the current project state into the agent's context. These commands are restricted to the project directory and are used for documentation analysis purposes.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the DESIGN.md file and source code in the internal/ directory, which could potentially contain malicious instructions.
Ingestion points: Contents of DESIGN.md and internal/ directory files.
Boundary markers: None present.
Capability inventory: Read-only operations including grep, sed, head, wc, git, and go list. The skill lacks file-write, network, or privilege escalation capabilities.
Sanitization: None present.

design-sync