The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill exhibits a vulnerability surface for indirect prompt injection because it reads and processes untrusted data from local files.
Ingestion points: Reads project content via read_file("README.md") and get_changed_files().
Boundary markers: Absent; there are no delimiters or instructions provided to the agent to ignore potentially malicious instructions embedded within the files it analyzes.
Capability inventory: Uses replace_string_in_file to modify the filesystem.
Sanitization: Absent; the skill does not appear to escape or validate the content retrieved from external files before performing updates.
[Data Exposure] (SAFE): The tool reads standard project documentation and directory structures. It does not target sensitive paths like SSH keys or environment secrets.
[Remote Code Execution] (SAFE): No remote scripts, package installations, or dynamic execution patterns were detected.

readme-updater