implementation-verify

Warn

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill metadata identifies a requirement for 'check-prerequisites.sh'. This file is missing from the skill package and originates from an untrusted repository (drillan/speckit-gates). This constitutes an unverifiable dependency that could execute malicious code.
  • PROMPT_INJECTION (LOW): The skill processes untrusted data from local project files. 1. Ingestion points: spec.md, tasks.md, and contract files. 2. Boundary markers: Absent. 3. Capability inventory: Parsing and generating fulfillment reports with recommendations. 4. Sanitization: Absent. Malicious instructions hidden in project requirements or task lists could influence the agent's report output or future implementation steps.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 10:51 PM