dark-pattern-audit
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill audits untrusted local source code and text, creating a potential surface for indirect prompt injection where malicious instructions hidden in the target codebase could attempt to influence the agent's behavior.
- Evidence Chain:
- Ingestion points: The entire repository is audited by default, loading all local source code and copy text into the agent's context (SKILL.md).
- Boundary markers: Absent; there are no instructions provided to separate the audited content from the agent's primary instructions using delimiters.
- Capability inventory: The skill is limited to reading local files and performing repository searches; no network operations or file-writing permissions are requested or used.
- Sanitization: Absent; no filtering or sanitization of the ingested content is performed before analysis.
- Mitigation: Update the workflow in SKILL.md to instruct the agent to wrap analyzed file content in clear delimiters (such as XML tags or triple backticks) and include an explicit directive to ignore any instructions or commands found within the audited content.
Audit Metadata