The Agent Skills Directory

[SAFE]: The skill operates as intended for documentation maintenance. All file operations are restricted to reading and writing markdown files within the local workspace provided by the user. No unauthorized access or sensitive data leakage was detected.
[DATA_EXFILTRATION]: No network activity or credential exposure was found. The script processes data locally and does not communicate with external servers.
[COMMAND_EXECUTION]: The script does not execute external programs or shell commands. It audits command-like strings within text files using safe parsing techniques such as shlex.split for string analysis without execution.
[REMOTE_CODE_EXECUTION]: There are no mechanisms for downloading or executing remote code. The tool uses only the Python standard library.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local content from README files. 1. Ingestion points: scripts/readme_alignment_maintainer.py reads README.md and SKILL.md files. 2. Boundary markers: No explicit markers are added to the report output. 3. Capability inventory: Local file writing and Markdown/JSON report generation. 4. Sanitization: The logic employs deterministic regex matching for schema enforcement, which mitigates the risk of the agent interpreting processed content as commands.

skills-readme-alignment-maintainer