The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use the Bash tool to execute directory creation commands (mkdir -p) using a path that can be optionally provided by the user. This represents a direct interaction with the system shell where unsanitized input could lead to directory traversal or command injection.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input and processes it into a file template without sanitization. 1) Ingestion points: The 'Trigger content' field in SKILL.md, which accepts arbitrary user input (coding snippets, categories, or thoughts). 2) Boundary markers: None; the skill explicitly states to 'use the user's original expression' and quotes it directly into the Markdown 'Trigger Material' section. 3) Capability inventory: The skill has the ability to create directories via Bash and write files to the local file system. 4) Sanitization: The instructions do not include any steps to escape or validate the user-supplied content before it is processed or written to disk.

analytic-memo