The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted data from external sources.\n
Ingestion points: The scripts/trac.py file retrieves ticket descriptions and comments from code.djangoproject.com, while scripts/forum.py fetches discussion posts from forum.djangoproject.com.\n
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat fetched data as untrusted or to ignore instructions embedded within the text.\n
Capability inventory: The skill has the ability to execute GitHub CLI commands via gh and perform filesystem operations including reading, searching (via grep), and writing triage reports.\n
Sanitization: Data is parsed using standard libraries and basic HTML stripping, but there is no mechanism to sanitize or filter potential prompt injection attacks hidden in user-generated text.\n- [EXTERNAL_DOWNLOADS]: The skill communicates with external domains to retrieve information.\n
Details: It fetches ticket metadata and RSS feeds from the official code.djangoproject.com domain and searches for discussions on forum.djangoproject.com using the Python standard library's urllib module.\n- [COMMAND_EXECUTION]: The skill executes external command-line tools and local scripts.\n
Details: It uses the authenticated gh CLI for searching and viewing pull requests on GitHub. It also executes internal Python scripts (scripts/trac.py and scripts/forum.py) and utilizes system tools like grep and glob to interact with a local clone of the Django source code.

django-ticket-triage