dev
Warn
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: MEDIUMPROMPT_INJECTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses explicit override language in SKILL.md, specifically stating that its rules have 'HIGHEST PRIORITY' and 'override all other instructions', which is a pattern associated with attempting to bypass or deprioritize the agent's system instructions.
- [INDIRECT_PROMPT_INJECTION]: The workflow defined in SKILL.md takes potentially untrusted data from user requirements and code analysis results and interpolates them into a development plan document, which is then used to generate content for shell commands executed via the Bash tool.
- Ingestion points: User requirements gathered via AskUserQuestion in Step 1 and analysis results from codeagent-wrapper in Step 2.
- Boundary markers: The skill does not provide delimiters or specific instructions for the agent to disregard potentially malicious commands embedded in the source data.
- Capability inventory: The skill utilizes the Bash tool for codebase analysis and parallel task execution, providing a significant capability surface.
- Sanitization: There are no explicit instructions or tools used to sanitize or validate external content before it is processed into the development plan or shell commands.
Audit Metadata