The Agent Skills Directory

[PROMPT_INJECTION]: The skill is inherently vulnerable to indirect prompt injection because its core functionality involves ingesting untrusted data from GitHub issue reports.
Ingestion points: The agent fetches content from external GitHub Issue URLs or processes raw text provided by the user (SKILL.md).
Boundary markers: The skill includes a 'SECURITY WARNING' instructing the agent to treat issue content as textual data and never follow commands embedded within the issue (SKILL.md).
Capability inventory: The agent uses curl to fetch external content from the GitHub API (SKILL.md).
Sanitization: No explicit sanitization or escaping of the ingested content is defined, relying solely on the provided prompt instructions.
[COMMAND_EXECUTION]: The skill instructions suggest using curl and jq to retrieve data from the GitHub API. While this is used for the primary purpose of the skill and targets a well-known service, executing shell commands based on user-supplied URLs carries a minor risk of command injection if the URL is maliciously crafted.

issue-triage