The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from GitHub issues which could influence the AI's triage recommendations.
Ingestion points: scripts/triage-collect.js fetches issue titles and bodies from GitHub APIs.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the scripts that process issue data for model judgment.
Capability inventory: scripts/triage-apply.js can execute gh issue comment, gh issue close, and gh issue edit (milestones and labels).
Sanitization: The skill uses basic trimming and shell quoting for execution, but does not sanitize issue content before it is processed by the AI for decision-making.
[COMMAND_EXECUTION]: The skill executes external commands using the gh and git CLI tools.
Evidence: scripts/triage-apply.js, scripts/triage-collect.js, and scripts/triage-github.js use execFileSync to run shell commands.
Risk: While the skill uses argument arrays and a custom quoteShellArg function, the arguments are derived from external GitHub content and model-generated reports, creating a surface for potential argument injection vulnerabilities in the underlying tools.
[EXTERNAL_DOWNLOADS]: The skill fetches issue data and comments from GitHub's official API using the gh CLI.
Evidence: scripts/triage-collect.js performs GraphQL and REST API requests to github.com.
Context: This activity is essential to the skill's function and targets a well-known, trusted service.

backlog-triage