planning-stories
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it ingests untrusted user input (story and feature descriptions) and interpolates it into generated markdown files. These files are intended to be processed by other subagents or autonomous scripts (the Ralph loop). The current templates do not employ explicit boundary markers or sanitization to prevent potentially malicious instructions within a user story from influencing downstream agents.
- Ingestion points: User-provided story descriptions and feature requirements captured via the Chat or AskUserQuestion tools (SKILL.md, ralph-tasks.md).
- Boundary markers: Absent; user input is directly inserted into designated sections of the STORY and TASK markdown templates without delimiters or 'ignore' instructions.
- Capability inventory: The skill utilizes the Write tool to create file system artifacts and has the capability to spawn a 'writing-tech-specs' subagent.
- Sanitization: No explicit sanitization or validation of the content provided by the user is performed before it is written to the planning documents.
Audit Metadata