izanami-product-writer

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests data from several untrusted external sources and interpolates that data into generation prompts without sufficient isolation.
  • Ingestion points: The skill reads data from app/specs/shared/project-spec.yaml, app/specs/account/subscription-spec.yaml, develop/account/authentication/func-spec.md, content/blog/posts/claudemix-qa.md, and .claude/skills/izanami-product-writer/docs/blog-catalog.md in prompts/01-collect.md.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when the extracted content is passed to the generation phases in prompts/02-generate.md.
  • Capability inventory: The skill leverages Read, Write, Edit, Glob, and Grep tools to interact with the filesystem.
  • Sanitization: There is no evidence of content validation or escaping for the data retrieved from the specification files before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 08:58 PM