The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input (the filename and week_id parameters) to perform file system operations and network requests.
Ingestion points: Untrusted data enters the agent context via the filename, week_id, title, and summary arguments in SKILL.md.
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for the file content being read or the data being processed.
Capability inventory: The skill possesses the capability to read local files (content from {filename}), write to the filesystem (published/{week_id}.json), and perform network operations (GET, POST, PATCH requests to a remote API).
Sanitization: There is no evidence of path validation for the filename or week_id inputs, nor sanitization of the file content before it is included in the network request body.

publish-weekly