The Agent Skills Directory

[Indirect Prompt Injection] (SAFE): The skill analyzes external data including markdown plan files and Rust source code. This represents a potential indirect prompt injection surface where malicious instructions in project files could influence the generated report. However, since this analysis is the primary intended purpose of the skill and it lacks capabilities to perform network exfiltration or high-privilege operations, the risk is classified as safe.
Ingestion points: plans/*.md, .rs files, Cargo.toml via find command.
Boundary markers: Absent; the skill relies on direct reading of these files.
Capability inventory: Limited to directory listing (find) and file reading for analysis.
Sanitization: Absent.
[Command Execution] (SAFE): The use of find is restricted to local directory traversal and is consistent with the skill's purpose of codebase inventory.

plan-gap-analysis