The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from external sources without explicit sanitization or boundary markers.
Ingestion points: Data enters the agent's context through the Gitee API calls list_repo_pulls, get_pull_detail, and list_comments as described in SKILL.md.
Boundary markers: The instructions do not define clear delimiters or warnings to treat the fetched PR content as potentially untrusted data.
Capability inventory: The skill uses Gitee API tools to read PR metadata and comments. It does not appear to have file-write, shell execution, or arbitrary network capabilities that could be leveraged for escalation.
Sanitization: There are no instructions for the agent to filter, escape, or sanitize PR titles, author names, or comment text before including them in the final report.

stale-pr-reminder