The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use various shell commands for filesystem manipulation.
Evidence includes: mkdir -p, cat >, mv, cp -r, rm -rf, find, and grep.
The commands are used to manage directories and files associated with AI agent tools (e.g., ~/.cursor/skills/, ~/.windsurf/rules/).
[DATA_EXPOSURE]: The skill provides the agent with access to sensitive configuration and instruction files for multiple AI developer tools.
It targets specific paths such as ~/.aider.conf.yml, ~/.github/copilot-instructions.md, ~/.codex/AGENTS.md, and ~/augment-guidelines.md.
These files define agent behavior and rules, and in some cases (like Aider), could contain sensitive operational metadata.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by design.
Ingestion points: The agent is instructed to read contents of SKILL.md and other configuration files using cat and grep across numerous tool directories.
Boundary markers: There are no instructions or boundary markers provided to ensure the agent disregards instructions contained within the files it is "managing."
Capability inventory: The skill possesses significant file system capabilities, including writing new files (cat >), moving files (mv), and deleting directories (rm -rf).
Sanitization: There is no evidence of sanitization or validation of the content being read from or written to these configuration files.

manage-skills