The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data, such as project source code and documentation, to identify content for consolidation or deletion. This ingestion process lacks explicit boundary markers or sanitization, creating an attack surface where instructions embedded within the analyzed data could potentially influence the agent's behavior.
Ingestion points: The skill analyzes project source code and documentation as specified in SKILL.md (Step 2) and references/evidence.md.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded in the processed content.
Capability inventory: The skill allows the agent to modify or delete files and execute git commands (SKILL.md, references/change-and-verify.md).
Sanitization: There are no instructions for sanitizing or escaping the content read from files before it is processed or used to make decisions.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to manage the refactoring process.
Evidence: references/change-and-verify.md requires running git status --short and git diff to identify workspace changes and verify modifications. While these are standard development operations related to the skill's primary purpose, they represent execution of system commands.

defrag