project-specification

Pass

Audited by Gen Agent Trust Hub on May 11, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill instructs the agent to automatically transition to the next phase of a project lifecycle (specifically attune:project-planning) without prompting the user for confirmation. This design is intended for workflow efficiency but bypasses standard interactive oversight gates.
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing external project briefs and using them to drive file operations and downstream skill triggers.
  • Ingestion points: SKILL.md (processes untrusted project briefs).
  • Boundary markers: Absent in the instructions.
  • Capability inventory: Writing to docs/specification.md and invoking Skill(attune:project-planning).
  • Sanitization: No sanitization or validation of the input content is described.
Audit Metadata
Risk Level
SAFE
Analyzed
May 11, 2026, 12:02 PM
Security Audit — agent-trust-hub — project-specification