The Agent Skills Directory

[SAFE]: The skill implements a comprehensive multi-phase analysis for inventory management, focusing on data discovery, par level optimization, waste tracking, and food cost analysis. These operations are performed using local data and do not involve unauthorized network communication or command execution.
[PROMPT_INJECTION]: The skill uses instructional markers such as "Do NOT ask the user questions. Analyze and act." to establish an autonomous agent persona. This is used for task-specific automation rather than bypassing platform safety guidelines.
[PROMPT_INJECTION]: The skill processes external data sources and user arguments, representing an indirect prompt injection surface.
Ingestion points: Processes user-provided $ARGUMENTS and local files identified during the discovery phase (e.g., POS logs, purchasing records).
Boundary markers: No explicit delimiters or boundary markers are used to encapsulate external data during processing.
Capability inventory: The skill is limited to reading local files, writing a markdown report to the docs/ directory, and appending execution metadata to local telemetry files.
Sanitization: No explicit sanitization or validation of ingested inventory data is specified.
[SAFE]: The skill includes a telemetry feature that appends execution results to a local file in ~/.claude/projects/. This access is restricted to recording outcome data and is consistent with project-level tracking within the Claude Code environment.

franchise-inventory