writing-flows

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill inherits a potential indirect prompt injection surface because it parses source code and specification documentation to map out architecture flows.
  • Ingestion points: Ingests project source files, handler scripts, and design specifications located via Glob specs/**/*.md using the Read and Grep tools.
  • Boundary markers: Absent. The instructions do not prescribe specific delimiters or instruction guardrails to isolate or ignore untrusted instruction strings embedded within code comments or text files.
  • Capability inventory: The skill possesses powerful system tools including Write, Bash, and SendMessage within its authorized tools configuration.
  • Sanitization: Absent. Content from code files and specs is consumed directly to generate flow files without preprocessing or structural validation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 08:08 PM