Skills
Navigate
skills/shinpr/codex-workflows/recipe-front-build/Gen Agent Trust Hub

recipe-front-build

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests data from local markdown files in docs/plans/tasks/ and docs/plans/ to drive sub-agent behavior, creating an indirect prompt injection surface. This risk is mitigated by the orchestrator's mandatory 4-step task cycle and final security-reviewer pass. Ingestion points: docs/plans/tasks/*.md and docs/plans/*.md in SKILL.md. Boundary markers: Absent. Capability inventory: File system modification (via sub-agents), git commit, and sub-agent orchestration. Sanitization: Absent.
  • [COMMAND_EXECUTION]: The orchestrator performs automated git commit and file cleanup operations after tasks pass quality gates. These actions are scoped to specific project directories and follow explicit user approval to enter autonomous mode.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 10:14 PM
Security Audit — agent-trust-hub — recipe-front-build