typo3-upgrade-effort-model

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides detailed workflows involving shell commands such as composer, php, grep, find, and jq to inspect project versions and code patterns for breaking changes. It also uses SQL queries to count database records for content elements, which is a standard procedure for upgrade estimation.
  • [EXTERNAL_DOWNLOADS]: Fetches extension metadata from Packagist (repo.packagist.org), a well-known and industry-standard repository for PHP packages. It also uses curl to retrieve the project's own sitemap for volume analysis, which represents legitimate diagnostic behavior.
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted project data during analysis. Ingestion points: Project source code (Classes/, Configuration/), composer.lock, package.json, and database record counts. Boundary markers: Absent; the agent is instructed to directly evaluate file contents. Capability inventory: Shell command execution (grep, find, composer, curl), file system reads, and SQL query execution. Sanitization: Absent; content from analyzed files is not explicitly sanitized. This surface is intrinsic to the skill's primary purpose of code assessment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 02:05 PM
Security Audit — agent-trust-hub — typo3-upgrade-effort-model