The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches documentation files from the vendor's GitHub repository (github.com/ashe-li/agent-skills) if local resource files are missing from the project directory.
[COMMAND_EXECUTION]: Uses shell commands via Bash to manage the development environment, including creating git worktrees (git worktree add) and running Python utilities for plan normalization.
[DYNAMIC_EXECUTION]: Executes a local Python script (plan_runner.py) from a dynamically resolved path ($HOME/Documents/agent-skills) to initialize, advance, and track the status of implementation plans.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied requirements and project source code to generate structured implementation plans that guide subsequent agent actions. This creates a vulnerability surface where malicious instructions embedded in the project files could influence the plan's logic.
Ingestion points: User input ($ARGUMENTS) and project source files accessed through Read, Glob, and Grep tools (SKILL.md).
Boundary markers: The skill uses a specific plan schema but lacks explicit delimiters or instructions to treat ingested project data as untrusted content.
Capability inventory: The skill utilizes powerful tools including Bash, Agent, Write, and Edit to execute the generated plans.
Sanitization: No specific sanitization, filtering, or validation of the ingested external data is described before it is processed by the planning agents.

design