generating-project-docs

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute local shell commands such as bun src/cli.ts, find, and git log to inventory repository assets and gather recent change context. These operations are restricted to the local filesystem for documentation purposes.
  • [PROMPT_INJECTION]: The skill ingests data from repository files which constitutes an indirect prompt injection surface.
  • Ingestion points: Processes content from package.json, SKILL.md frontmatter, and agent documentation files.
  • Boundary markers: No explicit delimiters are used for the ingested file content within the generation instructions.
  • Capability inventory: Includes shell command execution capabilities (bun, find) and the ability to modify project documentation files (README.md, ARCHITECTURE.md, STRUCTURE.md).
  • Sanitization: The skill contains explicit 'Quality Checks' requiring the agent to redact API keys, credentials, internal URLs, and local paths before saving the generated content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 07:58 PM
Security Audit — agent-trust-hub — generating-project-docs