create-feature-task
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via arguments and writes this content into a scaffolded markdown file in the
.claude/tasks/directory, creating a vulnerability surface where malicious input could influence future agent actions. - Ingestion points: The
<feature_name_and_description>argument defined in the frontmatter and used in Step 1. - Boundary markers: The markdown template in Step 3 lacks delimiters or instructions to treat the
{Feature Name},{Brief description}, and{captured}requirements as untrusted data. - Capability inventory: The skill writes files to the local file system and invokes the
TaskCreatetool. - Sanitization: There is no evidence of escaping, filtering, or validation of the user-supplied content before it is stored in the task documentation.
Audit Metadata