The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from GitHub (PRs, issues, and commits) which is user-generated and considered untrusted content.
Ingestion points: The skill retrieves data using tools like mcp_github_github_search_pull_requests and mcp_github_github_search_issues as described in references/workflow.md.
Boundary markers: While the skill uses structured Markdown and HTML templates for its reports, it does not explicitly define boundary markers or instruct the agent to disregard instructions embedded within the ingested GitHub data.
Capability inventory: The skill has file-writing capabilities (to .github/reviews/analytics/ and .a11y-history/) and performs command execution using node.
Sanitization: There is no evidence of sanitization, escaping, or filtering of the content retrieved from GitHub before it is rendered into report files.
[COMMAND_EXECUTION]: The skill initiates the execution of a local Node.js script for data processing and report generation.
Evidence: The Output contract in SKILL.md instructs the agent to execute node skills/a11y-core/scripts/render-report.mjs to complete the task.

analytics