The Agent Skills Directory

[COMMAND_EXECUTION]: The skill bundles a Python script scripts/detect_backend_framework.py designed to identify backend frameworks in a project by scanning manifest files. The script uses standard libraries and performs safe string analysis without executing external code.
[EXTERNAL_DOWNLOADS]: The SKILL.md file provides commands for the agent to install additional skills using a local helper script. This facilitates environment extension through defined management tools rather than arbitrary remote downloads.
[INDIRECT_PROMPT_INJECTION]: The project includes a framework detection script that reads untrusted manifest files. 1. Ingestion points: scripts/detect_backend_framework.py reads package.json, requirements.txt, and other project files. 2. Boundary markers: Absent. 3. Capability inventory: Subprocess calls are absent; the script only prints results to stdout. 4. Sanitization: Framework names are escaped via re.escape(), and versions are matched against a restricted alphanumeric regex.
[CREDENTIALS_UNSAFE]: Analysis of templates and examples shows adherence to secure credential handling. Documentation explicitly instructs users to manage secrets via environment variables and excludes them from version control, while implementation examples demonstrate the use of signed tokens for authentication.

backend