skills/versusjl/devin-blueprint/spec/Gen Agent Trust Hub

spec

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs file system operations to read project context (code, charters, notes) and write implementation specifications to a docs/ subdirectory.
  • [EXTERNAL_DOWNLOADS]: The skill is instructed to check official sources for current stable or LTS versions of dependencies and services mentioned in the specification.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by ingesting data from external, potentially untrusted sources.
  • Ingestion points: Reads referenced files, codebase, and Linear tickets (Workflow Step 1).
  • Boundary markers: No explicit markers or instructions to ignore embedded commands are used during data ingestion.
  • Capability inventory: File system writes and network operations for version checking and Linear ticket updates.
  • Sanitization: No input sanitization or validation of the ingested content is specified.
  • Mitigation: The risk is significantly reduced by a mandatory human-in-the-loop checkpoint ('Pause') that stops execution until the user reviews the generated spec.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 11:05 PM
Security Audit — agent-trust-hub — spec