tickets
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from feature specifications, project documentation, and source code to generate implementation tasks.
- Ingestion points: The skill specifically reads content from
ai/features/NN-feature-name/spec.md, various files indocumentations/, and project source code (file:SKILL.md). - Boundary markers: There are no explicit delimiters or "ignore embedded instructions" warnings used when processing the feature specification or documentation content.
- Capability inventory: The skill possesses local file-write capabilities (creating task files and roadmaps) and network publication capabilities (updating remote epics, tasks, and uploading ZIP archives if configured).
- Sanitization: The skill lacks sanitization or validation logic for the content found in specifications or documentation, which could lead to the agent following malicious instructions embedded in those files.
Audit Metadata