feature-prompt

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes grep and rg (ripgrep) to perform read-only searches within the codebase. This is used to identify inbound dependents and potential integration points for the requested feature.
  • [SAFE]: All file write operations are restricted to a specific documentation directory (docs/prompt/) and are used solely for saving generated artifacts. The skill instructions explicitly prohibit making edits to the existing source code.
  • [PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by ingesting data from repository files (such as CONTEXT.md and ADRs) and free-form user intake to generate prompts for downstream tasks.
  • Ingestion points: Processes project-level documentation files (CONTEXT.md, CONTEXT-MAP.md, docs/adr/*) and arbitrary user intake text.
  • Boundary markers: None identified; the skill does not explicitly use delimiters to wrap untrusted content in the generated output.
  • Capability inventory: Limited to read-only codebase scanning via shell commands and file-system write access within the documentation path.
  • Sanitization: No explicit sanitization or escaping is performed on the content retrieved from repository files or user input before interpolation into the output markdown.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 07:33 PM
Security Audit — agent-trust-hub — feature-prompt