triage-e2e-test

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on Node.js scripts that execute shell commands using execFileSync. These commands include git for repository state and ancestry checks, gh for GitHub API interactions, and unzip for log processing.
  • [EXTERNAL_DOWNLOADS]: The skill retrieves failure data from external sources including a Test Health API, Amazon S3 (for Playwright reports and logs), and GitHub PR search results. These operations are intended for gathering diagnostic information about test flakes.
  • [PROMPT_INJECTION]: The skill processes untrusted data from external sources, creating a potential surface for indirect prompt injection attacks.
  • Ingestion points: Test failure logs and reports fetched from S3 (scripts/fetch-pattern-evidence.js), diagnostic data from a history API (scripts/triage-history.js), and descriptions from GitHub PRs (scripts/find-prior-triage.js).
  • Boundary markers: Instructions require the agent to generate structured tables and read deterministic summaries (summary.md) rather than raw content when possible.
  • Capability inventory: The skill possesses capabilities for shell command execution (git, gh), local file system modification, and updating GitHub PR bodies.
  • Sanitization: The scripts include normalization logic for failure patterns and specific regex extractors for diagnosis fields, which provide a degree of sanitization for ingested content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 01:46 PM
Security Audit — agent-trust-hub — triage-e2e-test