skills/joacod/skills/project-specs/Gen Agent Trust Hub

project-specs

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because its core function involves processing and summarizing untrusted external data and repository files.
  • Ingestion points: The skill is instructed to gather and research information from 'mentioned URLs', 'user notes', and 'existing code' (SKILL.md, Steps 1 and 4).
  • Boundary markers: While the skill contains strong instructions to 'stop before build', it does not specify the use of delimiters or warnings to ignore instructions embedded within the research materials.
  • Capability inventory: The skill has the capability to write multiple documentation files to the file system (SKILL.md). It is explicitly restricted from executing production code or refactors.
  • Sanitization: No procedures for escaping or validating the content of external references or repository code are defined before they are incorporated into the project specifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:20 PM
Security Audit — agent-trust-hub — project-specs