agentic-qa-core

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documents workflows that ingest and process untrusted data from external sources like Jira (e.g., acceptance criteria and bug reports) to drive automated test generation and reporting.
  • Ingestion points: Jira issue content is synchronized to local markdown files via bun run jira:sync-issues for agent consumption.
  • Boundary markers: The references/skill-resolver.md protocol defines how to inject compact, authoritative rules into subagent briefings to limit the influence of untrusted data.
  • Capability inventory: The skill employs curl, acli, and various bun scripts for both read and write operations on the filesystem and external APIs.
  • Sanitization: references/jira-publishing-gotchas.md details the conversion process from Markdown to Atlassian Document Format (ADF) for secure data publishing.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute various CLI tools and scripts to perform environment validation, API testing, and Jira synchronization.
  • references/api-testing-doctrine.md specifies authenticated API execution using curl with sourced environment variables.
  • references/preflight-gate.md details the use of bun run commands for manifest checks and capability probing.
  • [EXTERNAL_DOWNLOADS]: The skill references the download and setup of development dependencies and testing tools from recognized providers.
  • references/preflight-gate.md mentions bun run pw:install for Playwright browser binaries and gentle-ai install for framework components.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 05:41 PM
Security Audit — agent-trust-hub — agentic-qa-core