The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes content from user-specified files, creating an indirect prompt injection surface. Maliciously crafted instructions within a target file could theoretically attempt to influence the agent's behavior during the optimization phases. However, the skill employs structured extraction and classification logic to mitigate this risk.
Ingestion points: The skill reads content from a user-provided file path in SKILL.md and phases/01-analysis.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present when processing file content.
Capability inventory: The skill has access to Write, Edit, and Bash tools, allowing it to modify files on the system.
Sanitization: Content is processed through classification logic but is not explicitly sanitized or escaped before being handled in the pipeline.
[COMMAND_EXECUTION]: The skill uses file-system tools to read and overwrite files. It implements a defensive design that extracts a functional inventory before optimization and performs an integrity check after modification. If the verification phase detects missing critical logic or new syntax errors, the skill is programmed to automatically revert the target file to its original state.

skill-simplify