skills/vllnt/skills/deliver-issue/Gen Agent Trust Hub

deliver-issue

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill is designed to execute various local commands to facilitate software development and testing.
  • references/vstack/capabilities/code-tests/REFERENCE.md specifies running all applicable required checks locally and recording commands and results.
  • references/vstack/capabilities/nextjs-tests/REFERENCE.md involves launching approved commands in managed sessions for server diagnostics and using browser automation tools like Playwright or Cypress.
  • references/vstack/capabilities/convex-components/REFERENCE.md includes executing codegen and test scripts.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection as it ingests untrusted external data and possesses high-privilege capabilities.
  • Ingestion points: Untrusted data enters the agent's context through issue descriptions, repository requests, and project configurations as described in the SKILL.md workflow.
  • Boundary markers: The skill refers to "applicable instructions and loaded principles" and uses explicit definitions of done, but lacks robust technical delimiters to prevent instruction hijacking from processed issues.
  • Capability inventory: The agent can perform file system writes (code implementation), create and update pull requests, and merge branches into target environments (SKILL.md).
  • Sanitization: The protocol in references/vstack/capabilities/nextjs-tests/REFERENCE.md requires findings to be supported by "sanitized evidence," and references/vstack/protocols/quality-validation.md mandates a reduction pass and independent review to justify code additions.
  • [EXTERNAL_DOWNLOADS]: The skill automates the installation of dependencies from the NPM registry to align with the vendor's stack.
  • references/vstack/capabilities/project-alignment/references/default-packages.md outlines a procedure to query the registry at runtime to identify and install the newest stable versions of packages within the @vllnt scope, such as @vllnt/logger and @vllnt/analytics.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 08:18 PM