task-verification-workflow

Pass

Audited by Gen Agent Trust Hub on Mar 18, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses authoritative directives like 'CRITICAL' and 'MANDATORY' to enforce specific agent workflows, overriding default task completion behavior to ensure deliverable quality.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it requires the agent to ingest and verify untrusted repository data.
  • Ingestion points: repository source files and progress logs (progress.txt).
  • Boundary markers: Lacks delimiters or instructions to ignore potential commands embedded within the files being analyzed.
  • Capability inventory: Utilizes browser automation tools (agent-browser) and shell commands (npm, npx, grep) to interact with the environment.
  • Sanitization: Does not perform validation or escaping of file content before it is processed by the agent.
  • [COMMAND_EXECUTION]: The workflow specifies the execution of local shell commands, including 'npm run dev', 'npx tsc', and 'grep', which are standard for code verification tasks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 18, 2026, 02:42 AM
Security Audit — agent-trust-hub — task-verification-workflow