The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's integration documentation in references/integrations.md instructs users to install and execute several third-party MCP servers using the npx -y command. While some sources are well-known, others refer to unverified community-maintained packages such as @sooperset/mcp-atlassian, @delorenj/mcp-server-trello, and @ibraheem4/linear-mcp. Executing unverified packages from individual contributors represents a risk of running malicious code.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it processes untrusted data from external project management platforms and user inputs.
Ingestion points: The skill retrieves data through MCP servers from platforms like Jira, Trello, Azure DevOps, Linear, GitHub, and GitLab, and also accepts manual input from users as described in SKILL.md and references/integrations.md.
Boundary markers: The instructions do not specify any delimiters or safety prompts to ensure the agent ignores malicious instructions embedded within issue descriptions or comments.
Capability inventory: The skill can query, create, and update issues and boards across various project management platforms, providing a significant impact surface if an injection occurs.
Sanitization: There is no evidence of sanitization, filtering, or validation logic for the content ingested from external tools or manual user summaries before it is used to influence the agent's decision-making process.

scrum-master