dpdp-agent-skill

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core functionality of auditing untrustworthy external codebases as defined in the workflows of SKILL.md. Malicious instructions embedded in audited files (such as code comments, database migrations, or schemas) could potentially influence the agent's behavior, particularly given its instructions to implement code changes.
  • Ingestion points: User-provided codebase files, database migrations, and logs analyzed in Workflow 1 of SKILL.md and the references/audit-checklist.md.
  • Boundary markers: There are no explicit instructions or delimiters defined to separate user data from agent instructions or to ignore commands within the audited content.
  • Capability inventory: The agent is empowered to perform file-writing operations and build compliance features as described in Workflow 2 of SKILL.md and the references/implementation-guide.md.
  • Sanitization: No specific filtering, validation, or escaping of the content of the analyzed files is prescribed before processing.
  • [NO_CODE]: The skill consists entirely of Markdown instructions, legal reference documentation, and templates. It does not include any executable scripts such as Python, JavaScript, or shell scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 11:09 PM
Security Audit — agent-trust-hub — dpdp-agent-skill