The Agent Skills Directory

[SAFE]: The skill provides a methodology and directory structure for managing project documentation (Context-Driven Development). No malicious patterns, obfuscation, or safety bypasses were detected in the instructions.
[COMMAND_EXECUTION]: The skill includes a local Node.js script (scripts/main.cjs) and execution hooks (hooks/pre-execute.cjs, hooks/post-execute.cjs). These scripts are boilerplate utilities for argument parsing and metric logging; they do not perform network operations, access sensitive credentials, or execute arbitrary shell commands.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to have the agent read and adhere to project-specific documentation stored in the .claude/context/ directory.
Ingestion points: Local markdown files including product.md, tech-stack.md, workflow.md, and files within the memory/ subdirectory.
Boundary markers: None; the methodology instructs the agent to treat these artifacts as the "single source of truth" for its behavior.
Capability inventory: The skill uses the Read, Write, Edit, Glob, and Grep tools to interact with the project context.
Sanitization: There is no automated validation or escaping of the natural language content stored in these files.

context-driven-development