The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/audit_repo.py script executes local git subcommands (rev-parse, ls-files, log) to gather repository metadata. These calls are implemented using subprocess.run with list-based arguments and without shell=True, which is the recommended safe practice for executing system commands.
[EXTERNAL_DOWNLOADS]: The documentation references official URLs from trusted platforms (openai.com, anthropic.com, opencode.ai). These are provided for informational purposes only; no remote scripts or executable binaries are downloaded or executed by the skill.
[CREDENTIALS_UNSAFE]: The skill includes a 'secret key' detection feature that uses regular expressions to identify potential hardcoded credentials in configuration files. However, the implementation is safe as it only flags the existence of the keys and explicitly avoids printing or processing the actual secret values.

agentic-readiness