make-project

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes external data from the repository environment to inform automated GraphQL actions.
  • Ingestion points: The skill reads CONTEXT.md, docs/agents/domain.md, and issue tracker configuration/metadata from the target repository.
  • Boundary markers: Absent. The instructions do not define delimiters or specific constraints to isolate the ingested data from the skill's execution logic.
  • Capability inventory: The skill executes several GitHub GraphQL mutations, including board creation, field/view configuration, and item population (create-board, link-repo, create-field, update-item-field).
  • Sanitization: Absent. There is no evidence of validation, filtering, or escaping of the ingested repository content before it is utilized in the GraphQL setup process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 03:00 AM
Security Audit — agent-trust-hub — make-project