find-duplicate-ownership
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and interpret untrusted data (source code) which may contain adversarial instructions aimed at influencing the audit outcome.
- Ingestion points: Files and symbols read across the codebase during the search and exploration phases defined in
SKILL.md. - Boundary markers: Absent. The prompt templates provided in
references/audit-prompts.mddo not include explicit delimiters or "ignore embedded instructions" warnings for the agent when processing external code. - Capability inventory: The skill employs multiple read-only subagents (
agents/duplicate-ownership-explorer.toml,agents/ownership-taxonomy-mapper.toml,agents/ssot-judge.toml) to scan, map, and judge codebase architectural patterns. - Sanitization: Absent. There is no evidence of content sanitization or escaping for the data retrieved from the files before it is interpolated into the agent's context.
Audit Metadata